--- Day changed

--- Log opened Mon Jun 16 00:00:01 2003
00:00 < alt> yeah..., I have win98se running on a P233-MMX, i'd rather have Linux but I have like one app I need to run
00:00 < alt> I just VNC in when I need it :)
00:01 < alt> but I figure this machine'll have at least 2 IRC servers on it, a webserver/mailserver and a DNS server all on seperate UMLs
00:01 < alt> I might run the web/mail on the machine itself tho
00:01 < Lathiat> bei interesting how a low-end server holds up in that environment
00:02 < Lathiat> im gonna plug this box in to try it, one monitor ;)
00:02 < Lathiat> bbs
00:03 < alt> well, if it's too low end, I suppose I can drop in my Compaq 1600 P3-550 
00:04 < alt> or this AMD K6-500 sitting next to me that I'm currently testing on
00:04 < alt> testing UML
00:06 < Lathiat> uh all
00:06 < Lathiat> doesnt power on
00:06 < alt> :(
00:07 < Lathiat> hrmz0r
01:51  * alt is away: bedtime
02:03 -!- shak [~shak@80.4.84.78] has quit [Remote host closed the connection]
02:11 -!- shak [~shak@80.4.84.78] has joined #uml
04:09 -!- shak [~shak@80.4.84.78] has quit [Quit: Client exiting]
04:55 -!- yeti [~yeti@p3EE2292C.dip.t-dialin.net] has quit [Quit: Client exiting]
04:58 -!- avoozzzl is now known as avoozl
05:14 < caker> linode.com is alive, I'm going to sleep
05:24 -!- yeti [~yeti@pD950B032.dip.t-dialin.net] has joined #uml
05:56 -!- DnsInfector [~DnsInfect@ARennes-204-1-4-167.w80-13.abo.wanadoo.fr] has joined #uml
06:02 < yeti> bonjour
06:02 < DnsInfector> hi
06:10 -!- shak [~shak@80.4.84.78] has joined #uml
06:20 -!- shak [~shak@80.4.84.78] has quit [Read error: Connection reset by peer]
06:42 < yeti> eeeeehhmm... a non-uml-question in between....   if i have an empty dir, create file1 and file2 and do a readdir, is there some rule that says the files have to be listed in that order?
06:42 < DnsInfector> depends on the underlying FS for me
06:42 < yeti> i think so too...
06:43 < yeti> would be strange otherwise, but humans are strange ;-)
06:45 < yeti>  
06:45 < yeti> next Q:
06:45 < yeti> you have experiences with umls in umls in skas mode?
06:45 < DnsInfector> umls ?
06:46 < DnsInfector> UML with ssl ?
06:46 < yeti> usermodelinux in usermodelinux running in skas mode instead of tt
06:46 < DnsInfector> yes
06:47 < DnsInfector> mine is running skas
06:47 < DnsInfector> just perfectly
06:47 < yeti> and you have usermodelinux kernels running ON an usermode system?
06:47 < DnsInfector> david confirmed me that the patch was 8 months long without a single problem
06:47 < yeti> i dont mean host with skas patch plus one level of uml above
06:48 < DnsInfector> ok
06:48 < yeti> i mean host with skas patch plus TWO level of umls above
06:48 < DnsInfector> i had understood
06:48 < yeti> ok
06:48 < DnsInfector> no i haven't tested such tricky case
06:48 < yeti> sometimes i think my english is bad
06:49 < yeti> so when i am not understood i retry more verbose
06:50 < yeti> i want to develop some uml managing software but i need to do some other things with my box too
06:50 < yeti> so i thought i do that inside an uml
06:50 < yeti> but that makes 2 lebels of umls...
06:50 < yeti> lebels->levels
06:51 < DnsInfector> i had the same idea
06:51 < DnsInfector> some kind of GUI over mconsole...
06:51 < green> Jeff developed the skas in exactly this way, I mean by running uml inside of uml
06:51 < yeti> i tried uml in uml maybe 10 months ago on gentoo and i had few luck with the networking.  maybe i just got confused by the routing
06:51 < DnsInfector> that would list UML running on the host
06:51 < yeti> green: good news
06:52 < DnsInfector> i also had problem 1 year ago with networking
06:52 < DnsInfector> but now everything is fine
06:52 < yeti> with one level of umls uml runs fine here for >>1 year now
06:54 < DnsInfector> I have build small GUIs with gtk1
06:54 < DnsInfector> and i wannted to testgtk2
06:54 < DnsInfector> perhaps i'll do something with mconsole to have a real project to perform
06:54 < yeti> i do many things in umls for >>1 year now
06:55 < yeti> its so nice to have systems for (maybe even destructive) experiments
06:55 < DnsInfector> :)
06:55 < DnsInfector> i used to test "software patch" in UML
06:56 < DnsInfector> ip-personnality
06:56 < DnsInfector> freeSwan
06:56 < yeti> fine... i'm nit sooooo fit in networking when it is more than ipv4... and guis are not my topic too...   now i know who to ask
06:56 < yeti> ;-)
06:56 < yeti> nit->not
06:57 < DnsInfector> :)
06:57 < yeti> i am using gnome atm, but i really feel at home on virtual consoles or inside screen
06:58 < yeti> all that colorfull bloat only confuses me
06:58 < yeti> i am at home in textmode but i try to get used to point'n'click
06:58 -!- radical [~kernet@210.7.75.194] has joined #uml
06:59 < yeti> pont'n'click means you cant access things you dont see...
06:59 < yeti> $PATH ist something different
06:59 < yeti> you can access things somewhere
07:00 < yeti> imagine you would have to cd to each application's dir before starting an application
07:00 < yeti> that would be kinda equivalent to crawling thru menus...
07:00 < yeti> i dont think current guis are state of the art
07:00 < DnsInfector> i do like console
07:01 < yeti> ;-)
07:01 < DnsInfector> but menu are a "good thing"^tm
07:01 < green> yeti: not necessary to go to each application's dir, you can just symlink all of them to your "desktop" ;)
07:01 < yeti> part-time-good(TM)
07:02 < yeti> i'd like to see much of the desktop functionality on shell level
07:02 < DnsInfector> i find that mconsole is quite poor in fonctionnalities
07:02 < yeti> like /mime/application/x-tar/extract my_file
07:02 < DnsInfector> i thought i could have grabbed much more informations on the running host
07:03 < yeti> can you throw an eye on the errors produced by readline if you pipe some text into mconsole?
07:03 < yeti> like        echo cad | uml_mconsole
07:04 < yeti> it is ugly
07:04 < yeti>  
07:04 < DnsInfector> is it me or mconsole is supposed to allow to activate Sys-rq keys
07:05 < DnsInfector> and SysRq are not present in the config menu....
07:05 < yeti> on kernel compile time?
07:05 < yeti> kernel-debugging subsection?
07:05 < DnsInfector> that's kernel hacking
07:05 < yeti> ups!
07:06 < DnsInfector> but i don't see sysrq
07:06 < yeti> just a moment!
07:06 < yeti>       [*]   Magic SysRq key
07:07 < yeti> it is in there
07:07 < yeti> but i looked at a vanilla kernel
07:20 < DnsInfector> 'm quite sure this option is in vanilla kernel
07:21 < yeti> just a moment, i'll look inside an uml kernel... need to unpack a tar first...
07:25 < yeti> un uml kernels it is in section general setup
07:25 < yeti> (2.4.20-3um)
07:53 < Lathiat> window 15
07:53 < Lathiat> erg
08:01 -!- radical [~kernet@210.7.75.194] has quit [Ping timeout: 490 seconds]
08:08 -!- shak [~shak@80.4.84.78] has joined #uml
08:09 < shak> hi
08:09 < yeti> ;-)
08:09 < yeti> hi!
08:09 < yeti> how's uml research?
08:09 < shak> I have 5 networked UMLs running :)
08:10 < yeti> using taps per uml or using uml_switch?
08:10 < shak> on this machine at the moment, Im just looking into making a nice filesystem with everything I need running
08:10 < shak> Im using taps per UML
08:10 < shak> it seems logical that way at the moment
08:10 < yeti> ok
08:11 < shak> the only question I have at the moment
08:11 < shak> is how can I make each UML have more bogomips (i.e. a more powerful CPU)
08:11 < yeti> you use skas mode?
08:11 < shak> I am using skas mode yes
08:12 < yeti> you have /tmp as tmpfs?
08:12 < shak> no, my /tmp is on disk
08:12 < yeti> thats it
08:12 < yeti> ok
08:12 < shak> so I need to have /tmp as tmpfs?
08:12 < green> shak: put more powerfult cpu in the box and you see more bogomips in uml ;)
08:13 < yeti> the uml kernel allicates a file which will be the uml's memory in tmp
08:13 < yeti> so if this is a ramdisk, it will give a speed boost
08:13 < yeti> but:
08:13 < yeti> i dont like that
08:13 < shak> green: I have a dual 1.5Ghz machine, which has 6000 bogomips, and each UML is using 1500 each
08:13 < green> yeti: not to the bogomips calculation. And after all, if there is enough memory in host system, even if /tmp is not memory-backed, still it will give no performance penalty
08:14 < yeti>  /tmp can be filled by every user
08:14 < yeti> i'd prefer to give each uml using user an own ramfisk
08:14 < green> shak: well, there is some misunderstanding, bogomips could not be used by somebody. those age BOGOMips ;_
08:15 < shak> well, I see bogomips as an indication of speed, say I wanted one UML to be more powerful than another one, is there anyway I can do that?
08:15 < green> shak: No.
08:15 < yeti> more ram to thr uml?
08:15 < green> shak: You can give one uml more cpu priority than the other one, though
08:15 < yeti> did you give some mem= argument to the uml kernel?
08:15 < shak> green: just by using "nice" and "renice"?
08:16 < shak> yes, I've given mem arguments to the kernels
08:16 < green> shak: yes
08:16 < yeti> jmmm
08:16 < yeti> check: is it the fs performance or the cpu speed that seems slow?
08:16 < shak> ok, I need to look at some way of limiting the amount of CPU one can use, because at the moment, if one UML decides to run something then it could use all the machine's resources
08:16 < yeti> i use uml filesystems on real partitions
08:17 < avoozl> i use cpu-cap to cap processor usage of uml processes
08:17 < avoozl> so i can run them at fixed 10% of real cpu for example
08:17 < shak> yeti: there isnt actually a problem with the speed, Im just curious
08:17 < avoozl> ofcourse it does distort the time inside the uml
08:17 < avoozl> but it does work quite well
08:17 < shak> avoozl: ah, that sounds ideal, thanks
08:17 < avoozl> shak: it does need a patch on the host kernel though
08:17 < yeti> avoozl: search the mailing list for fairsched or sth like that
08:17 < shak> avoozl: that's not a problem
08:17 < avoozl> http://www.tls-technologies.com/CPU/cpu-main.html
08:17 < yeti> but i have no experiences with that
08:19 < shak> hrm, it goes to 2.4.18
08:20  * shak tries to patch 2.4.21 with it and hopes
08:22 < shak> two hunks failed
08:24 < shak> so, either I can try and patch 2.4.18 up to date
08:24 < shak> or try and apply the hunks manually
08:24 < shak> avoozl: you using 2.4.18 then?
08:27 < avoozl> shak: 2.4.21-pre5-xfs-preempt-o1 or something
08:28 < shak> hrm..
08:28 < avoozl> i just fixed it manually
08:28 < shak> ok :0
08:28 < avoozl> i think the only thing that failed was some syscall table
08:28 < avoozl> you have to renumber the cap syscall i think, to a new free number
08:28 < shak> 1 out of 1 hunk FAILED -- saving rejects to file arch/i386/kernel/entry.S.rej
08:28 < shak> 1 out of 1 hunk FAILED -- saving rejects to file include/asm-i386/unistd.h.rej
08:29 < avoozl> yeah
08:31  * shak tries to get his head round this source
08:34 < shak> in this patch
08:34 < shak> in the .rej file
08:34 < shak> the first bit is the original is it?
08:36 < avoozl> > means added
08:36 < avoozl> < means removed
08:36 < avoozl> the rest is just context
08:37 < shak> ah 
08:37 < shak> thanks
08:40 < yeti> someone here uses mdz's debian packages of the uml kernel?
08:40 < yeti> on woody?
08:41 < yeti> when i try to rebuild them from source, compiling fails in some scsi stuff...
08:41 < shak> hrm, the first bit of this rej looks like it doesnt actually do anything
08:42 < shak> and the patch indicates that too
09:14 -!- shak [~shak@80.4.84.78] has quit [Quit: Client exiting]
09:21 -!- yeti [~yeti@pD950B032.dip.t-dialin.net] has left #uml [need to do some shopping or i'll stay with an near to empty fridge]
09:44 -!- braam [~braam@211.101.210.130] has joined #uml
09:45 < braam> hi
09:45 < braam> anybody any idea when we might see a UML for 2.5.71 (or 70)
09:48 -!- braam [~braam@211.101.210.130] has left #uml []
09:50 -!- mistral [mistral@212.159.71.212] has quit [Read error: Connection reset by peer]
09:52 -!- mistral [mistral@jstevenson.plus.com] has joined #uml
10:13 -!- DnsInfector [~DnsInfect@ARennes-204-1-4-167.w80-13.abo.wanadoo.fr] has quit [Ping timeout: 496 seconds]
10:20 -!- DnsInfector [~DnsInfect@ARennes-204-1-5-15.w80-15.abo.wanadoo.fr] has joined #uml
10:41 -!- shak [~shak@80.4.84.78] has joined #uml
10:45 < shak> hmm..
10:45 < shak> Checking that host ptys support SIGIO on close...No, enabling workaround
10:45 < shak> POSIX conformance testing by UNIFIX
10:45 < shak> when I reboot from inside the UML
10:45 < shak> it sits at that for ages
10:47 < shak> infact, it seems to have hung
10:48 < DnsInfector> so U're not posix compliant ;)
10:48 < shak> what can I do about that?
10:49 < shak> any any reason why it only does it when Im rebooting the UML?
10:50 < DnsInfector> no
10:50 < DnsInfector> i had troubles when "kill -9"ing my UML process
10:50 < DnsInfector> that's the only bug i've seen
10:51 < shak> you had problems after you'd done kill -9 or killall -9?
10:52 < shak> that could be the problem, I did have to killall -9 mine earlier
11:00 -!- yeti [~yeti@pD950B032.dip.t-dialin.net] has joined #uml
11:02 < DnsInfector> killall -9
11:02 < DnsInfector> i quite sure that there a deadlock in the code that handles tap devices
11:02 < DnsInfector> @ home i had uml_net which where waiting for tap device to die
11:02 < DnsInfector> and tap device which where wiating fort uml_net to release its references
11:02 < yeti> i only use preconfigured taps... long time on 2.4.19+skas3 host and never saw such a lockup
11:03 < yeti> now 2.4.21+skas3 but i upgraded yesterday so i cannot say much about .21+skas3
11:03 < shak> yeti, I was just commenting on the fact that when I reboot from inside the UML
11:03 < yeti> i have NO other patches inmy host kernel
11:03 < shak> I get the proble thatit pauses at POSIX conformance testing by UNIFIX
11:03 < shak> and hangs
11:04 < yeti> preconfigured tap or uml_net ?
11:04 < shak> uml_net
11:04 < DnsInfector> i forgot to mention that i was tcpdumping tap0 when all this happened
11:04 < DnsInfector> this could be the reason
11:04 < DnsInfector> uml_net
11:05 < yeti> hmmm... i can reboot my umls without problems but i use preconfigured taps a long time now...
11:06 < yeti> i cannot remember wether this was ok back when i used uml_net
11:06 < shak> maybe its worth me looking into them
11:06 < yeti> if you want hosting you want preconfigured taps too i think
11:06 < yeti> otherwise the uml owner could grab any address he wants
11:07 < shak> I'll check that one out now
11:10 < shak> are there any other filesystems other than the ones on the sourceforge site?
11:10 < shak> I need a better root_fs
11:11 < DnsInfector> shak: it's quite easy to doityourself
11:11 < shak> is mkrootfs the best?
11:12 < yeti> what do you need?
11:12 < yeti> better is no detailed information
11:13 < shak> well, there's three tools for making then
11:13 < yeti> roll your own
11:13 < shak> I was thinking that
11:13 < shak> is it a case of
11:13 < shak> I can install a distro here
11:13 < shak> boot into
11:13 < shak> configure it
11:13 < shak> and then use mkrootfs to make it usable for UML?
11:13 < yeti> forget that
11:14 < yeti> at least i never touched it and cannot say anything bout that
11:14 < shak> ok, thanks anyways
11:14 < yeti> how about starting with debian?
11:14 < shak> Im going to :)
11:14 < shak> I've got /tmp on tmpfs now
11:14 < yeti> debootstrap is said to run on other dists too
11:15 < shak> infact...
11:15 < shak> I dont seem to have tmpfs mounted on /tmp
11:15 < yeti> so basically you'd say "debootstrap woody woody a-debian-mirror
11:15 < shak> or, I could just use a spare partition
11:15 < yeti> you have anold system @work?
11:16 < shak> I have an old dual PII 400 next to me
11:16 < yeti> k6-2/500 still is fine for experiments
11:16 < yeti> when you put debian woody on that my scripts will run... ;-)
11:16 < shak> ;)
11:17 < yeti> how about putting woody on that one and i do the rest per ssh?
11:17 < yeti> *g*
11:18 < yeti> my stuff is so badly commented, installing it per ssh is easier than explaining it 
11:18 < shak> *G*
11:18 < shak> ok, you get your way, let me go download it
11:18 < yeti> how many harddisks?
11:18 < shak> its only a slow connection though
11:19 < shak> its just got one in the dual PII 400 at the moment
11:19 < yeti> everything above 4g should do for a while
11:19 < shak> its an 8gb disk
11:19 < yeti> minimal host system plus as much as possible for umls?
11:20 < shak> yep that's the way I'm going to try and do it
11:20 < shak> Im just going to download the disk
11:20 < shak> I'll probably do the install wednesday or something
11:20 < shak> because I have an exam tommorrow
11:20 < yeti> netinstall?
11:20 < shak> (I'm still a college student)
11:20 < shak> yep
11:21 < yeti> ok.  exams have higher prio
11:22 < yeti> think about a 2nd hd
11:22 < yeti> my "carrier" has a 4G system disk for the host and a 9G disk with logical volumes for the umls
11:23 < yeti> and i am thinking of ading more space for umls... maybe even one real hd per uml later...
11:23  * DnsInfector renice shak's exams
11:23 < shak> chuckle
11:23 < shak> its the last exam
11:24 < shak> I've had 25 exams already
11:24 < yeti> wow!
11:24 < shak> tommorrows is the very last one
11:24 < shak> just going to eat
11:24 < yeti> have fun and a bit of rest ;-)
11:25 < yeti> oftopic: how about linux and usb/bluetooth? ]:-)
11:37 -!- radical [~kernet@210.7.75.194] has joined #uml
12:13 -!- DnsInfector_ [~DnsInfect@ARennes-204-1-1-79.w193-251.abo.wanadoo.fr] has joined #uml
12:16 -!- DnsInfector [~DnsInfect@ARennes-204-1-5-15.w80-15.abo.wanadoo.fr] has quit [Read error: Connection reset by peer]
12:24 < shak> hmm
12:24 < shak> mkrootfs doesnt do debian does it?
12:35 < yeti> hmmm... when i want uml in uml with skas... does my 1st level uml kernel need the skas patch?
12:36 < shak> I think so
12:36 < yeti> hmmm...
12:36 < shak> I just trying the networking without uml_net
12:37 < yeti> i do the following...   each uml has an own uid/gid named uml#  with home in /home/uml#
12:37 -!- DnsInfector_ [~DnsInfect@ARennes-204-1-1-79.w193-251.abo.wanadoo.fr] has quit [Ping timeout: 492 seconds]
12:38 < yeti> i give each uml the tap with the same number
12:38 < yeti> the tap side gets 10.10.10.+ as address, the uml 10.10.11.#
12:38 < yeti> +->#
12:38 < shak> its more complex than the standard way
12:38 < yeti> invent some scheme like this to save a lot of thinking
12:38 < shak> and its not liking it here
12:39 < yeti> tunctl -u uml-userid -t tap#
12:39 < yeti> rule #1: if it is complex: script it
12:39 < shak> I've done that
12:39 < shak> what I just did
12:40 < yeti> ;-)
12:40 < shak> was tunctl
12:40 < shak> and then followed the HOWTO and added those bits
12:42 < david> shak: just use a bridge, rather than using two IPs for each UML
12:42 < shak> ok ;)
12:42  * shak goes off to read up about it
12:42 < shak> thanks
12:43 < BB> morning david 
12:48 -!- DnsInfector [~DnsInfect@ARennes-204-1-4-178.w80-13.abo.wanadoo.fr] has joined #uml
12:51 -!- shak [~shak@80.4.84.78] has quit [Quit: Client exiting]
12:53 -!- DnsInfector [~DnsInfect@ARennes-204-1-4-178.w80-13.abo.wanadoo.fr] has quit [Read error: Connection reset by peer]
12:55 -!- DnsInfector [~DnsInfect@ARennes-204-1-11-76.w81-53.abo.wanadoo.fr] has joined #uml
12:56 -!- shak [~shak@80.4.84.78] has joined #uml
13:00 -!- DnsInfector_ [~DnsInfect@ARennes-204-1-6-205.w81-49.abo.wanadoo.fr] has joined #uml
13:00 -!- DnsInfector [~DnsInfect@ARennes-204-1-11-76.w81-53.abo.wanadoo.fr] has quit [Read error: Connection reset by peer]
13:05 < shak> right
13:05 < shak> I seem to have found a bug
13:07 -!- shak [~shak@80.4.84.78] has quit [Remote host closed the connection]
13:09 -!- shak [~shak@80.4.84.78] has joined #uml
13:10 < david> shak: bug in what?
13:10 < shak> Im not sure whats wrong exactly, when I reboot any UML it pauses at the point when its checking POSIX compliance
13:10 -!- DnsInfector_ [~DnsInfect@ARennes-204-1-6-205.w81-49.abo.wanadoo.fr] has quit [Ping timeout: 492 seconds]
13:10 < shak> and hangs there
13:12 < shak> hm, that's funny
13:12 < shak> if I boot it without networking then it doesnt hang there
13:21 < shak> http://dev.uh-shells.co.uk/uml_log
13:21 < shak> there's a log of the system rebooting
13:24 -!- awyllie [~andrew@inhibition.dilex.net] has joined #uml
13:25 < shak> this is annoying.
13:25 < shak> I might post to the mailing list
13:49 -!- mistral [mistral@jstevenson.plus.com] has quit [Ping timeout: 492 seconds]
13:55 -!- mistral [mistral@212.159.71.212] has joined #uml
14:00 < shak> is there a guide anywhere for setting up a bridge for the UMLs?
14:07 < caker> Good afternoon
14:07  * alt is back (gone 12:15:26)
14:08 < shak> good afternoon
14:21 -!- shak [~shak@80.4.84.78] has quit [Quit: Client exiting]
14:41 -!- ASY- [~someone@216.221.51.54] has joined #uml
14:41 -!- ASY [~someone@216.221.51.54] has quit [Read error: Connection reset by peer]
14:41 -!- ASY- is now known as ASY
14:54 -!- avoozl [~jrk@vengeance.et.tudelft.nl] has quit [Ping timeout: 492 seconds]
15:07 -!- radical [~kernet@210.7.75.194] has quit [Remote host closed the connection]
15:12 -!- DnsInfector [~DnsInfect@ARennes-204-1-6-227.w81-49.abo.wanadoo.fr] has joined #uml
15:17 -!- radical [~kernet@210.7.75.194] has joined #uml
15:21 -!- DnsInfector [~DnsInfect@ARennes-204-1-6-227.w81-49.abo.wanadoo.fr] has quit [Ping timeout: 490 seconds]
15:32 -!- DnsInfector [~DnsInfect@81.53.247.77] has joined #uml
15:33 -!- radical [~kernet@210.7.75.194] has quit [Remote host closed the connection]
15:36 -!- mistral [mistral@212.159.71.212] has quit [Ping timeout: 488 seconds]
15:38 -!- radical [~kernet@210.7.75.194] has joined #uml
15:44 -!- shak [~shak@80.4.84.78] has joined #uml
15:44 -!- DnsInfector [~DnsInfect@81.53.247.77] has quit [Ping timeout: 488 seconds]
15:50 -!- mistral [mistral@jstevenson.plus.com] has joined #uml
15:50 -!- radical [~kernet@210.7.75.194] has quit [Remote host closed the connection]
15:53 -!- DnsInfector [~DnsInfect@ARennes-204-1-3-162.w193-252.abo.wanadoo.fr] has joined #uml
15:55 -!- radical [~kernet@210.7.75.194] has joined #uml
15:59 -!- shak [~shak@80.4.84.78] has quit [Quit: leaving]
16:02 -!- shak [shak@shell0.uh-hosting.co.uk] has joined #uml
16:03 -!- DnsInfector_ [~DnsInfect@ARennes-204-1-11-219.w81-53.abo.wanadoo.fr] has joined #uml
16:03 -!- radical [~kernet@210.7.75.194] has quit [Remote host closed the connection]
16:08 -!- DougD [~doug@68.109.90.156] has joined #uml
16:08 -!- DnsInfector__ [~DnsInfect@ARennes-204-1-9-182.w81-53.abo.wanadoo.fr] has joined #uml
16:09 < DougD> Does anyone here know anything about the UML community site new user registration process?
16:10 < caker> DougD: its busted
16:10 < david> DougD: yeah, it doesn't work :-)
16:10  * caker points at david
16:10  * BB points at david too
16:10 < david> DougD: I upgrade php-nuke and it stopped working - I'll look at it sometime
16:10 -!- DnsInfector [~DnsInfect@ARennes-204-1-3-162.w193-252.abo.wanadoo.fr] has quit [Ping timeout: 490 seconds]
16:11 < DougD> I suspected as much.  I have a commercial UML posting I would like to make, but would prefer not to be anonymous.
16:11 < david> DougD: mail it to me and I can post it to the site
16:11 < david> DougD: d@vidcoulson.com
16:12 < DougD> Thanks,  I will get it to you in a day or so.  We are a database hosting shop and have started using UML for in-house hosting functions.
16:12 -!- DnsInfector_ [~DnsInfect@ARennes-204-1-11-219.w81-53.abo.wanadoo.fr] has quit [Ping timeout: 492 seconds]
16:12 < david> DougD: okay - Maybe the site will be fixed by then :-)
16:12 < DougD> I will check tomorrow.  The symptom I am getting is that it sends me a confirm email, but when I click on it I get an account not found.
16:14 < shak> right, I'm starting with tmpfs for the /tmp instead of disk
16:14 < shak> I could do with more RAM
16:14 < caker> Woo!  First signup! 
16:14  * caker moshes around the room
16:16 < yeti> shak: i got my lazy ass up and converted my host to using a bridge
16:16 < shak> Im just about to do that 
16:16 < yeti> shak: the problem is the administration
16:16 < yeti> not the one time setup
16:17 < shak> david: you wrote the howto on uml,openconsultancy.com?
16:17 < yeti> from the beginning, think about the need to add umls without interferring others
16:18 < yeti> and i dont dare to use uml_switch up to now
16:18 < yeti> so each uml has a tap into the bridge
16:19 < Pahan> Ummm.
16:19 < Pahan> uml_switch is trivial.
16:19 < yeti> can the umls be run by different uids?
16:19 < yeti> can uml_switch be killed and restarted and the umls keep access?
16:19 < shak> anyway david, if you did, thanks for that, because I'm finding it really useful :)
16:20 < yeti> a real switch can be powerded off and on and the net will rerstart to work
16:25 -!- ElectricElf [david@elf.noc.oftc.net] has quit [Quit: Reboot.]
16:27 < yeti> can the bridged taps be filtered via iptables?
16:27 < yeti> in ipchains times that was an extra patch to the bridge code... but now?
16:28 < caker> yeti: you need ebtables and bridge-nf patches to get iptables to work
16:28 < shak> I'm just about to start now, I think I could do with more RAM for a tmpfs based solution
16:28 < yeti> caker: ok, then a topic for later...
16:29 -!- ElectricElf [david@elf.noc.oftc.net] has joined #uml
16:29 < david> shak: yes, I did :-)
16:30 < shak> well, thanks again for it :)
16:30 < shak> its great :)
16:30 < david> shak: your welcome
16:30 < shak> :D
16:30 < david> yeti: you can filter bridges with ebtables - You can filter tap devices with iptables
16:31 < Pahan> Hrm. "virtual memory files on /tmp"?
16:32 -!- radical [~kernet@210.7.75.194] has joined #uml
16:33 < david> yeti: if you restart uml_switch you have to delete then add the eth devices to the UML - Doable with uml_mconsole, but a pain in the ass
16:33 < david> uml_switch is a hack and I'd just use a bridge instead of it
16:33 < david> they pretty much do the same thing, but a bridge is more flexable
16:33 < yeti> i just converted to bridge
16:34 < yeti> instead of preconfigured taps + routing
16:34 < yeti> but i ned a way to prevent the uml owner to produce ip address conflicts
16:34 < david> yeti: so use iptables or ebtables to filter ARP responses to incorrect IP addresses
16:35 < yeti> i need to read about that...
16:35 < yeti> i need to filter the uml-side of the tap
16:36 < yeti> no malicious user shall harm his neighbours
16:36 < yeti> but i'll be busy on other things a while...
16:36 < david> yeti: to be honest, it's not a big deal - Large dedicated server hosting providers just deal with it
16:37 < david> yeti: You want to filter the host side of the tap device - Just tell it to block packets coming out the tap device (from the UML) with an incorrect IP address
16:37 < yeti> ok
16:37 < yeti> ;-)
16:37 < yeti> maybe ethernet and ip address as criteria... both have to be correct...
16:42 -!- radical [~kernet@210.7.75.194] has quit [Remote host closed the connection]
16:46 < shak> hmm..
16:46 < shak> yeti: did you patch your kernel for the bridging?
16:46 < shak> the host kernel that is
16:47 < yeti> no
16:47 < shak> because when I come to add a bridge I get br_add_bridge: Package not installed
16:47 < yeti> just vanilla kernel plus skas3 patch
16:47 < david> shak: modprobe bridge
16:47 < shak> I'll try that david, unless I dont have it in the kernel, which is most likely, I'm not really a modules person
16:48 < shak> modprobe: Can't locate module bridge
16:48 < david> shak: compile it in then
16:48 < david> CONFIG_NET_BRIDGE or somesuch
16:48 < shak> yep, another kernel recompile then :)
16:48 < shak> thanks :)
16:48 < yeti> ccache?
16:50 < shak> It appears just to be CONFIG_BRIDGE  :)
16:51 < david> shak: ah, well :-)
16:51 < david> shak: I was close
16:51 < david> shak: take a look and see what else you may need and compile that in too :-)
16:51 < david> saves a reboot later
16:51 < shak> I am doing :)
16:51 < shak> close but no cigar :p
17:10 -!- DnsInfector_ [~DnsInfect@ARennes-204-1-5-9.w80-15.abo.wanadoo.fr] has joined #uml
17:14 -!- DnsInfector [~DnsInfect@ARennes-204-1-2-112.w193-251.abo.wanadoo.fr] has joined #uml
17:16 -!- DnsInfector__ [~DnsInfect@ARennes-204-1-9-182.w81-53.abo.wanadoo.fr] has quit [Ping timeout: 496 seconds]
17:18 -!- DnsInfector_ [~DnsInfect@ARennes-204-1-5-9.w80-15.abo.wanadoo.fr] has quit [Ping timeout: 490 seconds]
17:21 -!- DnsInfector [~DnsInfect@ARennes-204-1-2-112.w193-251.abo.wanadoo.fr] has quit [Read error: Connection reset by peer]
17:40 -!- DougD [~doug@68.109.90.156] has left #uml []
19:34 -!- omphalion [~omphalion@68.43.95.100] has joined #uml
19:35 -!- omphalion [~omphalion@68.43.95.100] has left #uml []
19:56  * alt is away: Every time you masturbate... God kills a kitten. (Please think of the kittens)
19:56  * alt is back (gone 00:00:01)
20:20 -!- yeti [~yeti@pD950B032.dip.t-dialin.net] has quit [Quit: Client exiting]
22:08 -!- DnsInfector [~DnsInfect@ARennes-204-1-14-4.w81-248.abo.wanadoo.fr] has joined #uml
22:13 -!- ASY- [~someone@66.36.133.68] has joined #uml
22:15 -!- ruggy [~ruggy@qwest-png-ip123.scinternet.net] has joined #uml
22:19 -!- DnsInfector [~DnsInfect@ARennes-204-1-14-4.w81-248.abo.wanadoo.fr] has quit [Quit: Client exiting]
22:20 -!- ASY [~someone@216.221.51.54] has quit [Ping timeout: 488 seconds]
22:20 -!- ASY- is now known as ASY
22:30 -!- ruggy [~ruggy@qwest-png-ip123.scinternet.net] has left #uml []
22:47 -!- ichilton [~ian@pc3-stoc3-4-cust203.midd.cable.ntl.com] has quit [Remote host closed the connection]
22:47 -!- ichilton [~ian@pc3-stoc3-4-cust203.midd.cable.ntl.com] has joined #uml
23:30 -!- mistral [mistral@jstevenson.plus.com] has quit [Ping timeout: 492 seconds]
--- Log closed Tue Jun 17 00:00:00 2003