1148689192 Q * Methos Quit: Methos
1148689590 Q * yarihm Quit: Leaving
1148690724 J * doener_ ~doener@i5387C825.versanet.de
1148691137 Q * doener Ping timeout: 480 seconds
1148692489 J * DreamerC ~dreamerc@59-112-10-211.dynamic.hinet.net
1148692725 J * Methos ~mjoconr@hacker.pineview.net
1148693822 J * matti matti@linux.gentoo.pl
1148694286 Q * Naucki Quit: Verlassend
1148694311 Q * thom_ Quit: Verlassend
1148694871 M * Skram Hey all
1148694898 M * Skram i am running a gentoo base-- any slackware/debian/centos images floating around? pweese?
1148695027 M * doener_ you can use the debootstrap build method for debian guests, should work on any host distro as the necessary tools are downloaded by util-vserver
1148695034 J * id23 ~id@p54A0035F.dip0.t-ipconnect.de
1148695040 M * id23 hi #vserver
1148695045 M * doener_ hi id23 
1148695127 M * Skram doener_: how?
1148695234 M * doener_ vserver - build --help ;)
1148695239 M * Skram Guest images for Debian Woody & Sarge, Fedora, Mandrake, RedHat, Slackware and SuSE guests; all geared towards 2.4: http://debian.marlow.dk/vserver/guest
1148695272 M * Skram 2.4 :(
1148695281 M * doener_ http://linux-vserver.org/alpha+util-vserver -- section Vserver creation
1148695444 M * Skram Could not find local version of 'debootstrap'; downloading it from
1148695444 M * Skram http://ftp.debian.org/debian/pool/main/d/debootstrap/debootstrap_0.3.3_all.deb...
1148695447 M * Skram 21:03:52 URL:http://ftp.debian.org/debian/pool/main/d/debootstrap/debootstrap_0.3.3_all.deb [44962/44962] -> "/var/tmp/debootstrap.Hzv8sz/debootstrap.deb" [1]
1148695450 M * Skram I: Retrieving Release
1148695453 M * Skram E: Invalid Release file, no entry for main/binary-unknown/Packages
1148695455 M * Skram uhmm
1148695492 M * doener_ you didn't specify which version you want, eg. woody, sarge...
1148695512 M * Skram -d sarge
1148695521 M * Skram vserver dev build -m debootstrap --hostname dev --netdev eth0 --interface 70.86.176.49/26 --context 17649 -- -d sarge -m ftp://ftp.debian.org/debian 
1148695619 M * doener_ which arch is that? x86?
1148695629 M * Skram yes
1148695631 M * Skram its okay
1148695706 M * doener_ try adding --arch i386 
1148695752 M * Skram /usr/lib/util-vserver/vserver-build: unrecognized option `--arch'
1148695778 M * doener_ did you append that to the end of the command?
1148695784 M * Skram yeah
1148695785 M * Skram oh oops
1148695814 M * doener_ oops?
1148695822 M * Skram where should it have gone
1148695855 M * doener_ at the end should be fine... maybe it needs another -- in front (i don't remember which args are passed to which process)
1148695862 M * Skram okay
1148695867 M * Skram i did it at the begininng
1148695910 M * doener_ ...... -m ftp://foobar --arch i386 _or_ ...... -m ftp://foobar -- --arch i386
1148695915 M * doener_ one of these should work
1148696096 M * Skram to bring a debian box up to date..
1148696099 M * Skram apt-get update
1148696104 M * Skram and apt-get upgrade?
1148696170 M * doener_ yep
1148696180 M * doener_ i prefer using aptitude though
1148696190 M * Skram whats the difference?
1148696202 M * Skram GUI (ncurses)?
1148696333 M * doener_ yep, easy tracking of recommended/suggested packages, dependencies, searching, different views on the packages... quite a nice thing
1148696344 M * doener_ and yes, ncurses
1148696359 M * Skram 3.1 is woody
1148696361 M * Skram ?
1148696399 M * doener_ 3.0 is woody, 3.1 is sarge
1148696402 M * Skram since it is a VServer, I can remove /boot,floppy,initrd,and cdrom from /, right?
1148696406 M * Skram o ok
1148696408 M * Skram yeah
1148696457 M * doener_ yep, those aren't needed
1148696893 M * Skram just a quick refresher- in Debian, how does one set something to autostart at boot?
1148696982 Q * Methos Quit: Methos
1148697056 M * doener_ unlike gentoo, debian add most programs to the runlevels when you install them... to control runlevels, you can use update-rc.d
1148697065 M * Skram okay
1148697071 M * Skram i bet it added ssh :)
1148697142 M * doener_ update-rc.d has no -s option like rc-update, to check runlevels, I usually fall back to "ls /etc/rcX.d" (for runlevel X)
1148697150 M * Skram ok
1148697455 M * Skram http://linux-vserver.org/VServer+Hosting <-- how obnoxious! Lylix is a new company, and decided to take the liberty to put themselves at top.. if anyone else disagrees with their actions, feel free to move, Im  not going to however
1148697498 J * Methos ~mjoconr@hacker.pineview.net
1148697829 M * Skram doener_: what goes in the fstab of a RH9 vserver?
1148697879 M * doener_ inside the vserver or the one in the vserver configuration? the former shouldn't matter at all, the latter is probably just the same as for anything else... never used rh9
1148697897 M * Skram in the /etc/vservers/dev/fstab file
1148697917 M * Skram for debian and gentoo all ineed is:
1148697918 M * Skram none                     /proc                    proc     defaults             0 0
1148697922 M * Skram none                     /tmp                     tmpfs    size=16m,mode=1777   0 0
1148697925 M * Skram none                     /dev/pts                 devpts   gid=5,mode=620       0 0
1148697963 M * doener_ as I said, it's probably the same for rh9, but I've never used it, so I don't know if it has any special needs
1148697970 M * Skram ok
1148698141 Q * softi42 Ping timeout: 480 seconds
1148698597 M * doener_ hm, irc distracts me a lot today... back later ;)
1148698599 Q * doener_ Quit: leaving
1148698760 J * softi42 ciahqos@p549D74F8.dip.t-dialin.net
1148699367 Q * id23 Quit: Leaving
1148699819 M * Skram : )
1148703090 Q * Methos Quit: Methos
1148705769 J * Nam ~nam@S0106001195551ff0.va.shawcable.net
1148705859 M * Nam hey guys, I was wondering, I know that currently only 16 ip's can be assigned to a vserver, and there is a patch for 64, but what about going way more then that, is there a performance problem when you have more ip's setup, or would it be ok to custom patch it so you can setup way more then 64?
1148705905 M * cehteh you cna increase that number in the kernel, but it will slow down the system a bit
1148705968 M * Nam how does it effect the performance, and how much of a difference would it be with you have more?
1148705982 M * cehteh i once made a patch to allow masks for legal vserver ips which cost almost no performance
1148706011 M * cehteh planned to reinstantiate this thing, but i had no time/need for that yet
1148706012 M * Nam legal vserver ips?
1148706028 M * cehteh well the ip's a vserver an bind to
1148706038 M * Nam i've been using the old config where you assign it in the .conf file
1148706054 M * Nam i still need to convert to the new config
1148706083 M * Nam oh, i get ya, i read your comment wrong
1148706084 M * cehteh dunno how far ngnet is this days
1148706116 M * Nam i am building a hosting system with this so I would like to be able to support a lot of IP's
1148706162 M * Nam if all goes well, hopefully we will be able to not only promote this project, but donate some decent cash
1148706220 M * cehteh well first try to increase the limit in the source .. check the performance, if that is not goor for you, ask Bertl and/or me .. i would reinstantiate the patch fo a small consulting fee, but first i would acknowledge it with Bertl
1148706281 M * Nam coo, thanks
1148706326 M * Nam when i move over to our new server, i'm going to install everything updated, and i'll compile it with a larget limit and see how things go
1148706326 M * cehteh afaik is the IP check linear currently, and i dont know if it has to be checked for every packet but i guess so
1148706354 M * cehteh means on high traffic servers it might impact the performance seriously when you increase it
1148706386 M * cehteh but thats just an assumption, better try it out
1148706386 M * Nam hmm... I may just have to set it up with my load balance system then
1148706410 M * Nam have you herd of levanta? i think that's the spelling
1148706412 M * cehteh i guess you want to assign ips within some block to each server?
1148706431 M * Nam i'll probably assign each ip individually actually
1148706433 M * cehteh for loadbalancing ask derjohn 
1148706452 M * Nam because customers can rent the ip and it will grab a free one from the pool
1148706477 M * Nam i've already got some of that load balancing figured out, but thanks again, i may consult him
1148706484 M * Nam nice help in here ;)
1148706504 M * Nam always love an open source project which gives good irc support
1148706517 M * Nam your a developer on this project, or just do some side work?
1148706537 M * cehteh my masking patch wont help that much there
1148706594 M * Nam nope, probably not
1148706606 M * cehteh are your short on ip's, if not you could prereserve some blocks for each customer (4 8 16 ..) but just enable the ones he needs
1148706608 M * Nam but, i will be do alot of things, so i may, see how things go
1148706660 M * cehteh having 16 blocks a 16 ip's should prolly suffice :)
1148706660 M * Nam each time a customer needs an ip, a random unused one is grabbed from a pool
1148706690 M * Nam one thing i would like be able to do is add new ip without rebooting the vserver, is that possible?
1148706695 M * Nam i haven't looked into that yet
1148706703 M * cehteh mhm i think its better for routing when you block the ip's anyways
1148706741 M * cehteh dunno, but i think it could be made possible
1148706750 M * cehteh if it is not already
1148706765 M * Nam i am designing an autorouting system which is dynamic as I need it adjusted
1148706777 M * Nam kinda changing everything so it's more automated
1148706801 M * cehteh yeah sure
1148706826 M * cehteh most people do that with vservers :)
1148706906 M * Nam hehe, true, i'm just using the vservers as containers for all my servers, to get more use out of them of course, but that's only part of my system, I am automated the distribution process to all servers, in LAN or somewhere else, controls all configurations, auto routing, hardware detection/autoconfiguration, and a whole bunch more
1148706956 M * Nam right now, we have a really low load on the system, it's doing great
1148706975 A * cehteh is just software developer, not in hosting business ... well if i where i would like quite a lot of fancy features ... nested vservers and such :
1148706979 M * Nam 64bit athlon 3000+ with a gig of ddr400 and a raid sata drive
1148706983 M * Nam drives
1148707005 M * Nam nested vservers huh? i thought that didn't work
1148707011 M * Nam except with Xen
1148707021 M * cehteh yes .. just thinking about
1148707051 M * Nam i doubt i'll have a need for nested vservers, but it would be interesting
1148707057 M * cehteh well in some way it works .. you can configure the vservers to be started from the root .. but nest their filesystems
1148707080 M * Nam what I would really like, is a way to run the same vserver on two different systems, so it responds and helps do some load balancing
1148707085 M * Skram hey all
1148707093 M * cehteh there are quite a many nice things which could be done
1148707108 M * Nam hey Skram
1148707114 M * Skram HiHI
1148707125 M * cehteh :)
1148707135 M * Nam is that possible cehteh?
1148707169 M * cehteh real nesting unfortunally not .. just the filesystems, sure
1148707185 M * Nam no, i ment running the same vserver on two different systems at the same time
1148707195 M * Nam on xen apparently you can do that
1148707201 M * Nam haven't tried xen out though yet
1148707258 M * cehteh we had a failover solution at the linuxtag ... with distributed block devices
1148707299 M * cehteh running 2 exact vservers at the same time is more a problem of what you are running and how you want to syncronize the data
1148707334 M * cehteh at least writes need to be serialized/locked somehow .. 
1148707366 M * Nam yea, hmm... I wonder how xen does it, of it i misread something
1148707371 M * cehteh and you need some shared filesystem, blockdevice, database, ... whatever
1148707384 M * cehteh no xean has only hot migration
1148707395 M * cehteh if i am not wrong
1148707403 M * Nam like if you wanted to run mysql, it locks files, you couldn't have it writing over eachother like that
1148707411 M * Nam or many other programs
1148707416 M * Nam pid lock files and stuff
1148707416 M * cehteh means you can transfer a xen machine to another host without rebooting
1148707465 M * cehteh which is somewhat useless in some cases since you have no automatic failover when a machine dies .. it just servers for scheduled maintainace 
1148707467 M * Nam yea, i don't know how that would work, bcause like, it would be different ip's and stuff, you would have to reconfigure the ip's
1148707502 M * Nam the failover I was designing was pretty simple
1148707511 M * Nam run the vservers on a really high quality NAS
1148707517 M * cehteh another funny thing which once worked was that you can patch vserver+openmosix .. means you have a frontend vserver and many nodes in the back
1148707526 M * Nam with redundant hotswappable PSU/HDs
1148707540 M * cehteh but i dont know if that still works, neither would i suggest to use that in production
1148707553 M * Nam then just run two servers, if one goes down, the other starts the vservers  on it
1148707583 M * Nam hehe, that would be interesting
1148707600 M * Nam i ran an openmosix cluster once
1148707602 M * Nam in my house
1148707609 M * Nam a bunch of old systems
1148707620 M * cehteh i am not profund in load balancing ... prolly i would just run a frontend squid and some web/database servers in the back
1148707621 M * Skram our company is working on a console-based control panel for VServer
1148707621 M * Skram :|
1148707649 M * Hollow morning
1148707654 M * cehteh moin Hollow 
1148707660 M * Skram hey Hollow 
1148707678 M * Nam i have a console based ones as well, i'm going a little farther then that though
1148707695 M * Nam check this out
1148707695 M * Nam http://www.levanta.com/
1148707701 M * cehteh there is this new web frontend for vserver :)
1148707705 M * Nam this is a smaller product
1148707714 M * Nam mine is going to be far superior to this
1148707724 M * Nam it's php/dhtml/ajax my version
1148707782 M * Nam hehe, this is a cool part, I provide a system which allows sites to be made which interact with all the control panel features, so a customer can create their own control panels build into their sites
1148707931 M * cehteh .. so coffee ...
1148707938 M * cehteh brb
1148707942 M * Skram wtf
1148707948 M * Skram ill stop development then
1148707948 M * Skram lol
1148707972 M * Skram where is the VServer part?
1148707994 M * Nam ?
1148708908 M * Skram Levanta isnt free..
1148708910 M * Skram pfft
1148708919 M * Skram Nam: what "system"?
1148709180 M * Nam hehe, it's not public yet
1148709188 M * Nam within the year we will be though
1148709212 M * Skram Need help with development?
1148709215 M * Nam the system is a advanced Content Managenet System
1148709234 M * Skram Oh- Right on
1148709243 M * Nam kinda... at the moment we need to get things going before we bring on more people
1148709249 M * Skram Whats Laventa have to do VServer
1148709252 M * Nam the dept is costing more then it's bringing in right now
1148709270 M * Nam you watch their video, that's what they use... well, a server virtualization
1148709378 M * Hollow Nam: how do you interact with the host the vservers are running on?
1148709520 M * Nam i have built it all in php, and i use command line to interact with the vservers, i had to custom program PHP with c++ to extend it's abilities to get it to do it, because currently it PHP doesn't have the ability to do it
1148709547 M * Nam i've writen all custom api's for controlling everything
1148709555 M * Nam i have based it off of slackware/slamd64
1148709555 M * Skram nice
1148709575 M * Nam it's been a big project
1148709575 M * Hollow so you call the util-vserver commands from within your php scripts
1148709582 M * Nam been working on it for 2.5 years
1148709589 M * Nam yup
1148709627 M * Nam for remote server, I tunnel it all through ssh
1148709634 M * Nam slower but secure
1148709659 M * Hollow ic.. do you want a xmlrpc interface? :)
1148709683 M * Nam i haven't worked with xml at all so far, haven't really liked it that much
1148709696 M * Skram *drools*
1148709698 M * Nam eventually though, i will be forced to have it
1148709699 M * Hollow well, you don't have to know xml, to call methods via xmlrpc
1148709702 M * matti Hollow: :))
1148709710 M * Hollow matti: *g*
1148709715 M * matti :>
1148709731 M * Hollow Nam: it's already there
1148709760 M * Hollow apropos.. anyone of you willing to test the daemon?
1148709774 M * Hollow i'll try to prepare some test images & scripts next week
1148710591 J * bubulak ~bubulak@cicka.wnet.sk
1148714423 Q * s0undt3ch Ping timeout: 481 seconds
1148714688 J * Methos ~mjoconr@hacker.pineview.net
1148714889 J * bonbons ~bonbons@83.222.38.81
1148715838 J * Viper0482 ~Viper0482@p54977469.dip.t-dialin.net
1148716073 M * jake- morning
1148716573 Q * shedi Read error: Connection reset by peer
1148717343 J * shedi ~siggi@inferno.lhi.is
1148718133 M * tokkee Morning.
1148719671 J * ntrs ~ntrs@62.162.191.4
1148719693 Q * ntrs Quit: 
1148719997 J * DarthVader ~Aniken@203.177.212.163
1148720331 Q * bubulak Ping timeout: 480 seconds
1148721726 J * bubulak ~bubulak@cicka.wnet.sk
1148721967 J * yarihm ~yarihm@84-74-17-70.dclient.hispeed.ch
1148722014 Q * Methos Quit: Methos
1148722858 Q * gerrit Ping timeout: 480 seconds
1148723652 J * Methos ~mjoconr@hacker.pineview.net
1148723915 P * Methos 
1148724670 J * dna ~naucki@p54BCE3C5.dip.t-dialin.net
1148725286 J * dna_ ~naucki@p54BCFDCF.dip.t-dialin.net
1148725402 J * Naucki ~naucki@p54BCDCDA.dip.t-dialin.net
1148725653 Q * dna Ping timeout: 480 seconds
1148725747 M * derjohn cehteh, issue solved?
1148725757 M * cehteh ?
1148725762 M * cehteh vom query?
1148725773 Q * dna_ Ping timeout: 480 seconds
1148725792 M * cehteh .. bin noch am überlegen
1148726005 J * dna_ ~naucki@p54BCDCDA.dip.t-dialin.net
1148726016 J * Dr4g debian-tor@tor-irc.dnsbl.oftc.net
1148726433 Q * Naucki Ping timeout: 480 seconds
1148726594 M * mire I just love it takes only a few processes to run a distro :)
1148726598 M * mire http://rafb.net/paste/results/Qy6GHE84.html
1148726623 M * Hollow gpm? ;)
1148726629 M * mire yeah, will kill that ;)
1148726801 M * mire err, vim and ncurses require gpm...can't kill it... stupid deps
1148726824 J * dna ~naucki@p54BCDCDA.dip.t-dialin.net
1148726894 Q * dna Quit: 
1148727068 Q * dna_ Ping timeout: 480 seconds
1148728581 J * simon_ ~simon@212.76.255.43
1148729123 Q * DarthVader Quit: Leaving
1148729148 J * DarthVader ~Aniken@203.177.212.163
1148730696 J * the_hydra ~a_mulyadi@202.59.168.5
1148730960 J * doener ~doener@i5387C825.versanet.de
1148731551 J * s0undt3ch ~s0undt3ch@bl7-245-140.dsl.telepac.pt
1148732092 Q * the_hydra Quit: 
1148732542 M * s0undt3ch can I share a device witha guest?
1148732625 M * s0undt3ch I have /dev/authfail to listen to f_authpriv for ssh sscript kiddies, it does it's job for the servers, but as for a guest that is "uploading" it's syslog messages to my main server syslog nothing happens
1148732917 M * s0undt3ch never mind a simple syslog-ng change made it work
1148734121 Q * simon_ Remote host closed the connection
1148736186 Q * nebuchadnezzar Ping timeout: 480 seconds
1148737723 N * Bertl_oO Bertl
1148737727 M * Bertl morning folks!
1148737780 M * Bertl s0undt3ch: nevertheless, to answer that question: yes, you can for most devices .. but most of them are dangerous to give to a guest
1148737830 M * Bertl mire: but you can stop the service, IIRC
1148737841 M * Bertl mire: s/stop/disable/
1148740151 M * daniel_hozac bonbons: you seem to have broken the IPv4 address hiding ;)
1148740195 M * bonbons daniel_hozac: Oh, should not... what IPv4 addresses do you see that you shouldn't?
1148740202 M * daniel_hozac all of them.
1148740203 Q * mire Quit: Leaving
1148740230 M * bonbons when I tested I saw only those that I should have seen... do you have IPv6 modular or not, module loaded?
1148740252 M * Bertl daniel_hozac: hide_netif is on?
1148740258 M * daniel_hozac damn, no.
1148740276 M * daniel_hozac although, the IPv6 address hiding is enabled even without that.
1148740297 M * Bertl well, not a big deal, I ahve to work through it anyways ...
1148740308 M * Bertl the important thing is that it works properly for ipv6
1148740481 M * bonbons the hiding should be happening at same time as for IPv4... daniel_hozac: where do you get the wrong output? (e.g. what command shows wrong data)
1148740505 M * daniel_hozac ./chbind6 -4 127.0.0.1/8 -6 1111::3/120 ip a ls
1148740544 M * daniel_hozac hides all IPv6 addresses except 1111::3.
1148740970 M * Bertl bonbons: may I ask a few questions to your patch?
1148740987 M * Bertl bonbons: this is the latest version, yes? patch-2.6.16.18-vs2.1.1rc21-ipv6a.diff
1148741117 M * tokkee Is the logo (http://vserver.13thfloor.at/Stuff/LOGO/Linux-VServer-1.02.eps) also available as .fig?
1148741198 M * Bertl hmm, I guess not, can ps be converted to .fig?
1148741237 Q * mkhl Quit: 
1148741238 M * Bertl ah, yes, pstoedit, sec
1148741261 M * tokkee Bertl: Cool, thx.
1148741354 M * Bertl have to download and compile it, but should be simple ...
1148741403 M * tokkee I already have it installed... just a minute.
1148741433 M * Bertl ah, good, I can put the conversion results there if you upload them somewhere
1148741483 M * tokkee Hmmm... the result is not very nice :-(
1148741516 M * Bertl what are your plans with the .fig?
1148741530 M * tokkee Include it in a LaTeX document.
1148741541 M * Bertl hmm, eps is simpler, isn't it?
1148741581 M * Bertl http://amath.colorado.edu/documentation/LaTeX/reference/figures.html
1148741614 M * tokkee Yes... after converting it to pdf its fine, too...
1148741634 M * tokkee I was just trying to have all input files as fig ;-)
1148741642 M * tokkee s/its/it's/ ;_)
1148741648 M * tokkee s/_/-/ 
1148741657 M * Bertl ah, i.c. :)
1148743111 M * bonbons Bertl: yep, it's the last version
1148743300 J * `DiaMonD ircap751@201.246.120.18
1148743310 M * Bertl welcome `DiaMonD!
1148743320 M * `DiaMonD thank
1148743322 M * `DiaMonD :D
1148743393 M * Bertl bonbons: patch is here: http://www.13thfloor.at/~herbert/patch-2.6.16.18-vs2.1.1rc21-ipv6a.diff.hl, comments/questions are here: http://www.13thfloor.at/~herbert/ipv6.txt
1148743414 J * nebuchadnezzar ~nebu@zion.asgardr.info
1148743422 M * `DiaMonD [Bertl] i need psy bnc
1148743431 M * `DiaMonD help me please
1148743433 M * `DiaMonD :$
1148743442 M * Bertl what is psy bnc?
1148743470 M * `DiaMonD eMm sorry but my english not perfec
1148743471 M * jake- a bouncer
1148743480 M * daniel_hozac hmm, isn't struct in6_addr a uint8_t[16]?
1148743484 M * `DiaMonD i´m from chile in southamerica
1148743497 M * Bertl daniel_hozac: no idea, didn't have a look :)
1148743513 M * Bertl daniel_hozac: if so, then it's fine for me :)
1148743517 M * `DiaMonD ok thank u
1148743531 M * daniel_hozac in6_addr{ union { __u8[16], __u16[8], __u32[4] } }
1148743548 M * `DiaMonD [jake-] you know psybnc?
1148743550 M * jake- vs~Bertl: http://en.wikipedia.org/wiki/Bounce_(network)
1148743555 M * jake- sur
1148743557 M * jake- e
1148743565 J * thom_ ~thomas@chello080108114160.11.11.univie.teleweb.at
1148743574 M * Bertl welcome nebuchadnezzar! thom_!
1148743593 M * nebuchadnezzar hi
1148743605 M * bonbons Bertl: relating to IPv6 module unloading, don't know when one can do so, after modprobing it has a refcount of 8 I can't decrease (even vanilla)
1148743643 M * nebuchadnezzar bonbons: it's normal according to USAGI guys
1148743645 M * Bertl bonbons: okay, do you see any use in having it modular?
1148743679 M * Bertl what I mean is, if we drop the ipv6 as module option, we can avoid 60% of the changes no?
1148743720 M * bonbons the use is to use IPv6 only when needed (not that useful for server I assume), it's especially defining when one starts using IPv6
1148743752 M * bonbons no, the modular IPv6 part is relatively small
1148743773 M * Bertl doesn't look to me that way ...
1148743855 M * tokkee http://tokkee.org/~tokkee/tmp/linux-vserver-arch.pdf ... What do you guys think about that?
1148743862 M * bonbons it's a few lines at 3 places: addition to _init functions, it's just for the filtering of IPv6 devices, the rest is modular-independent
1148743893 M * Bertl hunk 5, major part of ifa_in_nx_info changes, cleanup_ipv6_mibs
1148743959 M * Bertl (basically all CONFIG_IPV6_MODULE checks)
1148743996 M * Bertl OTOH, what is the advantage of ipv6 being modular when you cannot unload it?
1148744009 M * bonbons everything where CONFIG_IPV6 and CONFIG_IPV6_MODULE are separate
1148744024 M * Bertl it's highly unlikely that you say, hey today I want to start doing ipv6 :)
1148744046 M * bonbons you can load it when you start using it, useful for laptops ;)
1148744067 M * Bertl does it harm the laptop when it is compiled in?
1148744133 M * tokkee Probably not, but there some people who do not like to have stuff in their kernel as long as they don't really need it :-/
1148744160 M * Bertl well, those folks (like me) will then _not_ compile it into the kernel
1148744166 M * bonbons not really. Only drawback is that the laptop is usually not silent with IPv6 (ICMPv6 router discovery)
1148744178 A * h01ger moins
1148744182 M * h01ger jake-, is Jim Knopf really your realname? (i want to put it in the thanks file - and i know a fictional character Jim Knopf, thats why i ask :) sorry for that..
1148744192 M * Bertl bonbons: how do you get it silent again, once ipv6 is loaded?
1148744253 M * bonbons through sysctl I think, but not sure at all (never tried that, even don't know how to ask kernel to resent discovery request)
1148744332 M * Bertl can ipv4 be modular?
1148744454 M * Bertl doesn't look like ...
1148744473 M * Bertl so personally I'd prefer to force ipv6 to be on or off
1148744498 M * Bertl and to avoid all the indirections and racey checks
1148744525 M * bonbons no, IPv4 cannot be module it's all or nothing
1148744530 M * jake- holger: of course not :)
1148744666 M * bonbons Bertl: for all IPv4 part, does it make sense to replace uint32 by in_addr? This way we are sure we get things right on all arches (big-endian/little-endian) [related to NIPQUAD/atoquad comment and equiv for in6_addr]
1148744697 M * daniel_hozac bonbons: seems a guest can't bind to :::x if the host has bound <host IPv4 address>:x (sshd in this case)
1148744755 M * Bertl bonbons: in the userspace/kernel ABI we want strict types like uint* in the kernel we can use in_addr and friends
1148744757 M * bonbons daniel_hozac: that's possible, I don't do remapping...
1148744758 M * daniel_hozac however, after stopping the host's sshd, starting the guest's, and then restarting the host's, they both seem to work fine.
1148744780 M * Bertl that's the missing address conflict v6
1148744810 M * bonbons Bertl: yep
1148744880 M * bonbons Bertl: what are your intentions for remapping (lo) and userspace API changes?
1148744927 M * Bertl let me put the lo patch there too .. sec
1148745028 M * daniel_hozac re: endianness, aren't IP addresses always MSB?
1148745031 M * Bertl http://www.13thfloor.at/~herbert/delta-lo0.05.1.diff.hl
1148745052 M * Bertl daniel_hozac: yes, at least on the network
1148745071 M * Bertl daniel_hozac: but AFAIK throughout the kernel as well, see our IPI_LOOPBACK
1148745107 M * Bertl #define IPI_LOOPBACK htonl(INADDR_LOOPBACK)
1148745170 M * daniel_hozac right.
1148745192 M * Bertl bonbons: regarding remapping, we should not need to remap either 0.0.0.0 or 127.0.0.1 in the future .. we only have to choose a proper source ip from the set of available ips
1148745219 M * Bertl bonbons: regarding API/ABI, either we add a special command (add/remove/etc) for ipv6 addresses
1148745235 M * Bertl or we bump the version to v1 and make it understand/use the union
1148745257 M * Bertl here for me the important questions are related to the mask/prefix issues
1148745282 M * Bertl - do we need an ipv4 mask at all? or would a prefix do nicely?
1148745298 M * Bertl - does ipv6 know network masks? or just prefixes?
1148745351 M * Bertl besides the direct match, we probably want to add more advanced versions of address matching, like:
1148745362 M * Bertl - match to ip and broadcast
1148745367 M * Bertl - match to entire network
1148745401 M * Bertl maybe even some kind of hierarchy or algebra like:
1148745409 M * Bertl - add this network
1148745416 M * Bertl - remove that network
1148745426 M * Bertl - union/intersection or such?
1148745451 M * Bertl because all those changes are much faster than the linear comparison
1148745459 M * Bertl s/changes/checks/
1148746125 M * daniel_hozac using a prefix for IPv4 means 2 extra calculations in ip_find_src (only place where it's actually used, AFAICT).
1148746190 M * Bertl well, doesn't mean we have to store it as prefix in the kernel
1148746221 M * Bertl how does ipv6 handle the prefix? and is there something like a non-prefix mask at all=
1148746225 M * Bertl s/=/?
1148746447 J * stefani ~stefani@c-24-19-46-211.hsd1.wa.comcast.net
1148746531 M * Bertl wb stefani!
1148746543 M * stefani hola
1148746607 M * matti Bertl: :)
1148746689 M * bonbons from my knowledge, IPv6 uses only prefix, not masks, and this mostly for routing purposes (/64, /128, /48 not much more in use)
1148746722 M * Bertl hey matti!
1148746774 M * matti ;-)
1148746781 M * matti Jesus Christ.
1148746808 M * matti I recived 83. IP pool from RIPE.
1148746853 M * matti And almost from out-of-box I am blocked in more or less ten RBL databases ;]
1148746857 M * matti Huh...
1148746975 M * matti :)
1148747234 M * Bertl hmm, sounds like configuration issues :)
1148747286 Q * DreamerC Ping timeout: 480 seconds
1148747310 M * bonbons does it make sense to pass over prefix/mask when giving the guest an IP address? I would even say that the mask configured on the host side the the one of interest
1148747335 M * Bertl ahem, well, yes and no ... depends on for what the mask is used ...
1148747384 M * Bertl for simple ip checks, the mask is not of any use
1148747409 M * Bertl for all other checks, the mask might make sense, even if it differs from the host mask
1148747458 M * bonbons do you have an example?
1148747466 M * Bertl further, the mask set on the host (interface) can have a completely different meaning than for the network checks, but it should be the one used for broadcast checks and broadcasts
1148747483 M * Bertl sure, just assume we configure the host interface to:
1148747490 M * Bertl 10.0.0.1/16
1148747517 M * Bertl the guest gets the ips 10.0.1.1-10.0.1.15
1148747544 M * Bertl we could write/send that as 10.0.1.0/28
1148747565 M * Bertl but that doesn't mean that broadcasts would use 10.0.1.15
1148747571 M * matti :-)
1148747603 M * Bertl (actually I would have to either add 10.0.1.0 or remove 10.0.1.15 :)
1148747679 M * bonbons for IPv6 broacast does not exist anyhow :)
1148747725 J * DreamerC ~dreamerc@59-112-13-159.dynamic.hinet.net
1148747745 M * bonbons for IPv4 it would then be better to add addresses as such or as range + adding broadcast addresses
1148747987 Q * DarthVader Quit: Leaving
1148748561 Q * DreamerC Quit: leaving
1148748578 J * DreamerC ~dreamerc@59-112-13-159.dynamic.hinet.net
1148748836 M * Bertl hmm, ipv6 does not know broadcast or multicast?
1148748862 M * Bertl how do you ping a network in ipv6?
1148748866 M * bonbons IPv6 only know multicast
1148748881 M * bonbons to ping you send (multicast) ICMPv6
1148748927 M * Bertl hmm, could you give an example for a network?
1148749001 M * Bertl i.e. the equivalent for: ping -b 10.0.0.0 e.g.
1148749168 M * bonbons pinging networks is not possible! You can ping hosts or multicast targets (like allrouters, allhosts, ...)
1148749207 M * Bertl aha, interesting ...
1148749262 M * bonbons ping6 -I eth0 ff02::1 (ping all nodes), ff02::2 (all routers), ff02:3 (all hosts)
1148749358 M * bonbons nodes are any IPv6 machines on the local link, routers are what they are ;) and hosts should be DNS boxes (not completely sure on this one)
1148749523 M * bonbons broadcast+multicast (IPv4) has been simplified to multicast only (IPv6)
1148749585 M * bonbons if you took a look at IPv6 addresses you get be default, there are a few types: multicast addresses (type determined based on the first block)
1148749759 M * Bertl sounds reasonable .. judging from the mess broadcast causes
1148749783 M * bonbons then there are link local addresses (e.g. fe80:* the interface gets automatically even with unplugged cable) and globally routable addresses like 2001:* you might have seen in AAAA dns answers
1148749864 M * bonbons in general you have /48 or /64 prefixes for networks, the 64 last bits determining the host inside of it's network. (thos 64 bits are generated based on MAC address unless privacy option is enabled)
1148750218 M * Bertl but /49 would be fine too? or is that forbidden?
1148750283 M * bonbons that would be fine too, you can set whatever you want, but that's only used for routing anyhow or maybe autogeneration of address in network
1148750367 M * mnemoc hi!, does vserver-utils have a homepage?
1148750370 M * bonbons normally in IPv6 you have a router advertisment daemon that sends out ICMPv6 messages saying "I'm router for you and can offer you network with prefix 2001:*/48 or 2001:*/64" (or at your will something between those 2)
1148750381 M * daniel_hozac mnemoc: http://dev.croup.de/proj/vserver-utils
1148750391 M * mnemoc daniel_hozac: thanks
1148750412 M * bonbons mnemoc: yep (but partially out of date): http://dev.croup.de/proj/vserver-utils/
1148750551 M * mnemoc i have been moving my vservers to tap devices quite succesfully these days, what do you think about making it by default if the defined device doesn't exisit? :)
1148750567 M * mnemoc instead of fooling around prepre-start and postpost-stop
1148750581 Q * matti Ping timeout: 480 seconds
1148750594 M * Bertl mnemoc: are you sure about that?
1148750611 M * mnemoc Bertl: what side efects do you see?
1148750632 M * Bertl no, what I mean is, are you sure you are _using_ tap devices?
1148750641 M * mnemoc uhm
1148750661 M * mnemoc at least ip addr shows the ips bindinded to the tap devices :(
1148750666 J * matti matti@linux.gentoo.pl
1148750670 M * Bertl check the packet counts there
1148750686 M * mnemoc uhm
1148750764 M * mnemoc aren'T they counted?
1148750764 M * Bertl and once you figured that, please let me know what the difference to using dummy device(s) would be
1148750780 A * mnemoc puzzled :(
1148750904 M * mnemoc Bertl: why is it 0 ? :(
1148750934 M * Bertl so, instead of messing with pre/post start/stop you can simply add a 'dev' entry with  "dummy0" and use that for all guests 
1148750948 M * Bertl (assumed that you did load the dummy device)
1148750967 M * Bertl mnemoc: it is 0 because the device is not used in your case
1148750977 M * doener Linux networking uses IP address and interface separately. It doesn't matter on which interface the addresse is defined, routing always considers all interfaces
1148750994 M * doener s/addresse/address/
1148750995 M * Bertl mnemoc: you would have to attach some kind of userspace process to the tun/tap and setup routing there to actually use it
1148751052 M * mnemoc my goal was just to not bind to the external (and only) interface, but dummy does the same :(
1148751075 M * mnemoc :(
1148752640 P * stefani parting (is such sweet sorrow)
1148753440 J * DarthVader ~Aniken@203.177.212.163
1148754829 M * Bertl welcome DarthVader!
1148756488 J * tat_ ~caganet@98.Red-83-58-124.dynamicIP.rima-tde.net
1148756547 M * tat_ can someone explain me how to use "resource sharing" on vservers, or point me to a document , i couln't find anything on the web
1148756608 M * daniel_hozac for sharing what?
1148756677 M * tat_ files that probably wont alter lik /bin /sbin ...
1148756687 M * daniel_hozac so unification?
1148756693 M * tat_ yes.
1148756694 M * daniel_hozac http://linux-vserver.org/alpha+util-vserver
1148756747 M * tat_ is alpha the port, or does it mean unification is in alpha ?
1148756760 M * daniel_hozac huh?
1148756797 M * Hollow apropos.. dows CoW obsolete vunify and/or vhasify?
1148756824 M * doener Hollow: hm, you still need something to apply CoW to ;)
1148756845 M * Hollow yeah, a template/base image is still needed
1148756850 M * Hollow but no more hash calc and such, right?
1148756863 M * daniel_hozac well, reunification after updates?
1148756920 M * Hollow mhm..
1148756981 J * IrcAp-C ircap751@201.246.120.18
1148756984 M * doener CoW makes exclude-lists/logic unneeded, and you can just create a full shallow copy of the template, but as daniel said, vservers change and you might want to reunify
1148756986 M * Hollow maybe we can discuss requirements for vcd wrt unification, i'm still a noob on this topic ;)
1148757034 M * daniel_hozac is that really something you need the daemon to do?
1148757049 M * daniel_hozac IMHO it's more of a cron job like task.
1148757068 M * Hollow yeah, that's the question, what vcd has to do (if anything)
1148757087 M * Hollow i'm more thinking in the direction of vserver creation
1148757094 M * Hollow which is handled by the daemon
1148757143 M * daniel_hozac cp -al?
1148757170 M * Hollow ;)
1148757357 Q * `DiaMonD Ping timeout: 480 seconds
1148757368 J * `DiaMonD ircap751@201.246.120.18
1148757497 Q * IrcAp-C Ping timeout: 480 seconds
1148757847 M * bonbons Hollow: regarding unification you have a few samples in vserver-utils-1.0.4 branch ;) (cowcp and vunify; simple versions but effective)
1148757894 M * Hollow yeah, will take a look at it
1148757909 M * Hollow bonbons: btw, can you provide an init-ng image?
1148759237 M * Skram hercules dev # rsync -av --progress /vservers/ /nas/1/vservers/
1148759238 M * Skram HAHA
1148759243 M * Skram this is going to take a while
1148759253 M * Skram will probably error out, huh?
1148759268 J * dd ~ircap751@201.246.120.18
1148759447 Q * `DiaMonD Ping timeout: 480 seconds
1148759596 M * bonbons Hollow: that should be possible, how much do you want inside of the image?
1148759630 M * Skram i have a couple vpses that all belong to me (same server), and I want to unify parts of them.. how?
1148759676 M * Skram would i need to rebuild the vservers? restart?
1148759678 M * Hollow bonbons: doesn't really matter, mabay one or two services to test (sshd & nc e.g.)
1148759696 M * Skram : )
1148759718 M * bonbons Hollow: ok, let me kook that together, I have already most of it :) (my vs-tiny)
1148759730 M * bonbons s/kook/cook/
1148759737 M * dd please help me
1148759738 M * dd :$
1148759741 M * dd i need a server to install psybnc on
1148759888 M * daniel_hozac Skram: you'd need to restart them before you'd see the memory benefits of shared binaries/libraries/etc.
1148759894 M * Skram daniel_hozac: okay
1148759903 M * Skram so how do i see how much it will unify/
1148759923 M * Skram dd: look into a VPS provider which allows it
1148759952 M * daniel_hozac Skram: vserver ... hashify -nv
1148759977 M * dd thank :(
1148759990 M * Bertl okay, off for now .. back later ...
1148759995 N * Bertl Bertl_oO
1148760010 M * Skram daniel_hozac: ... is where i put a VServer name
1148760014 M * Skram hrmph
1148760049 Q * thom_ Quit: Verlassend
1148760091 M * Skram Failed to initialize unification for vserver
1148760128 M * Skram im using util-vserver.. is it (v)unify?
1148760168 M * daniel_hozac you probably didn't set it up for unification.
1148760176 M * Skram right
1148760182 M * Skram there are configs or what?
1148760195 M * daniel_hozac well, not really, just creating a directory should suffice.
1148760220 M * Skram a directory, meaning /vserver/NAME ?
1148760226 M * Skram eh
1148760243 M * daniel_hozac no, more like /etc/vservers/.../apps/vunify.
1148760248 M * Skram ok ok
1148760265 M * doener  http://linux-vserver.org/alpha+util-vserver
1148760286 M * Skram hercules sentien-mail01 # vserver sentien-mail01 hashify -nv
1148760286 M * Skram Initializing exclude-list for /vservers/sentien-mail01 (sentien-mail01)
1148760286 M * Skram Can not determine packagemanagement style
1148760286 M * Skram failed to determine configfiles
1148760289 M * doener there's a section on vhashify that explains the necessary setup
1148760290 M * Skram yucky but ok
1148760532 Q * tat_ Quit: BitchX-1.1-final -- just do it.
1148761275 Q * eyck Quit: leaving
1148761686 J * eyck eyck@ghost.anime.pl
1148762492 Q * eyck Quit: leaving
1148762953 Q * DarthVader Remote host closed the connection
1148763065 J * eyck eyck@ghost.anime.pl
1148764620 M * bonbons Hollow: vs-tiny is being uploaded to my homepage (http://homepage.internet.lu/vserver), it includes ssh, metalog, core-utils, initng, nc (patched 0.6.4 version)
1148764762 M * bonbons to start it with initng you need to use /sbin/initng as init, the root password for it is vs-tiny
1148765093 M * bonbons uploaded. The initng patch is to get initng's output working inside of vserver (some unknown issue with stdout...). It's configured to use /dev/tty12 as /dev/console, sshd and metalog are in runlevel
1148765144 M * bonbons ups typo in homepage! Should have been http://homepage.internet.lu/brunop/vserver/ !
1148768937 Q * cohan Ping timeout: 480 seconds
1148768940 Q * blizz Ping timeout: 480 seconds
1148769054 Q * Dr4g Remote host closed the connection
1148769167 J * CoBoLt ~countcobo@d54C498B4.access.telenet.be
1148769170 M * CoBoLt oi guys
1148769176 M * CoBoLt just a small question
1148769184 M * CoBoLt have you tested the patches on an alpha?
1148769197 M * CoBoLt as I am getting problem compiling my kernel on a miata...
1148769206 M * CoBoLt error goes with the osf_sys.c
1148769219 M * CoBoLt the kernel builds fine without the patches
1148769223 M * CoBoLt but once patched
1148769227 M * CoBoLt no go anymore :(
1148769242 M * CoBoLt I am running CentOS for alpha on it
1148769255 M * CoBoLt wanted to compile the latest kernel... Patching didn't give any errors
1148769296 M * CoBoLt I will be afc for a couple of hours (it's 00:36 am here and I'll go sleeping now
1148769300 M * CoBoLt I'll come back tomorrow
1148769396 Q * sid3windr Ping timeout: 480 seconds
1148769421 J * sid3windr luser@bastard-operator.from-hell.be
1148769515 Q * Viper0482 Quit: bin raus, 
1148771634 M * ray6 ~.
1148771648 M * ray6 aeh ja
1148771720 M * doener i totally agree ;)
1148771847 Q * derjohn Remote host closed the connection
1148771894 M * Hollow bonbons: thanks!
1148771955 M * bonbons just note that it's all compiled for pentium-m so it won't work on too old CPUs
1148772009 M * Hollow ok, hope it will work on athlon-xp ;)
1148772030 M * bonbons it should, that one has SSE no?
1148772039 M * Hollow yup
1148772085 M * bonbons and even if not, most should work, there's not too much floatingpoint in a basic system!
1148772090 M * doener what about sse2?
1148772119 M * Hollow no sse2
1148772127 M * Hollow at least /proc/cpuinfo says that ;)
1148772136 M * doener my pentium-m has sse2
1148772179 M * Hollow we'll see
1148772194 M * bonbons mine have too, but I don't think this will hurt too much. If a binary fails, just replace it with one of your host system (it's all gentoo ;))
1148772239 M * Hollow hehe.. i will prepare some more images for different init styles, and some test scripts next week, so we can test the basic things
1148772268 M * bonbons that's a good idea!
1148772315 M * bonbons the best thing would be to have the script to generate the images ;) (though there are such things at releng no?)
1148772349 M * Hollow well, you can use catalyst for that..
1148772384 M * Hollow but i also need some other distro for the sysvinit style
1148772412 M * bonbons is there a good manual for catalyst somewhere?
1148772490 M * Hollow well, some old doc is around at http://www.gentoo.org/proj/en/releng/catalyst/
1148772514 M * Hollow you should get the idea, but i assume some config options have changed meanwhile
1148772585 M * Hollow you can also take a look at http://dev.croup.de/proj/gentoo-vps/browser/stages/build-stages it simplfies the creation of our stages
1148772599 M * Hollow (but it can be used for every stage)
1148772604 M * bonbons is it also possible to update catalyst images?
1148772655 M * Hollow well, catalyst keeps binpkgs, so it will probably rebuild the whole image but will use the previous binpkgs for that
1148772690 M * bonbons my aim would be kind of minimal base system with packages added binary-wise (binpkg) on top (filtering out all dev-files)
1148772695 J * Methos ~mjoconr@hacker.pineview.net
1148772784 M * Hollow mhm.. i.e. you don't want a compiler and things inside right?
1148773056 J * derjohn ~derjohn@80.69.37.19
1148773111 M * Hollow hey derjohn!
1148773170 M * derjohn Hey, Hollow ... ! BTW: Did you pass your exams?
1148773195 M * Hollow don't know yet.. have still one on tuesday too
1148773211 M * Hollow results will be known after pfingstferien ;)
1148773240 M * derjohn tuesday? *thumbs*
1148773261 J * cohan ~cohan@koniczek.de
1148773278 M * Hollow physics.. *bah*
1148773279 M * Hollow ;)
1148773313 M * Hollow really looking forward to our "work retreat" hope it will work out
1148773355 M * derjohn well, I didnt hear from it yet. Are there any updates? maybe we should vote for a date ... 
1148773374 J * blizz ~blizz@evilhackerdu.de
1148773376 M * Hollow heh, thought so too, no activity on the page for a while
1148773381 M * derjohn (of course BenBen and Cemil are the tie-breakers, as it's their room)
1148773382 M * Hollow maybe we should announce it on the ml too
1148773391 M * Hollow sure..
1148773406 M * derjohn I think the announcement should be after we found a date.
1148773426 M * Hollow well, some may not have noticed the idea yet
1148773446 M * bonbons Hollow: yep, and if possible limit need for perl/python as well!
1148773449 M * derjohn else, we'll not find a date that fits to all.
1148773484 M * Hollow bonbons: some time ago i stumbled across BINPKG_HOST, so it should work out if you provide binpkgs on another host
1148773528 M * Hollow derjohn: indeed
1148773542 M * bonbons and then there was also the $ROOT for portage (just don't know where /var/db/pkg is maintained under that scenario)
1148773598 M * derjohn Hollow, i think the best way is to force BenBen amd cemil to offer dates ...
1148773599 M * Hollow one thing i also thought about was: we really need some info page about distro maintainers, with contact and links 
1148773617 M * Hollow bonbons: maintained under the root
1148773640 M * bonbons which root? the one where working portage is installed or $ROOT?
1148773648 M * Hollow $ROOT
1148773700 M * Hollow derjohn: yup, good idea ;)
1148773980 M * derjohn Hollow, e-mail sent and I sent /myself now to S1 or S3 mode .... leaving office ... n8 !
1148773991 M * Hollow night!