1106266234 M * Bertl hmm?
1106266283 M * narender1983_ait why do i need a devpty in my vserver
1106266314 M * Bertl to open new ptys?
1106266386 M * narender1983_ait what if i dont need an X do i still need devpts
1106266413 M * Bertl ptys are used for all kind of stuff, like screen for example
1106266448 M * narender1983_ait so there presence is mandatory
1106266502 M * Bertl for a working system, yes
1106266524 M * narender1983_ait that solved one of my problems
1106266580 M * narender1983_ait secondly how r access to filesystem handeled in a vserver
1106266608 M * Bertl the same as on the host ...
1106266633 M * narender1983_ait means no special checkings
1106266662 M * Bertl except for the unification and barrier stuff ... no
1106266666 M * narender1983_ait like xid tags etc as is written in the paper on 13thfloor.at
1106266697 M * narender1983_ait could please explain how is that done
1106266795 M * Bertl the tagging? or what?
1106266819 M * narender1983_ait no unification and barrier
1106266999 M * Bertl well, both are flags, extending the basic functionality
1106267074 M * narender1983_ait why do we need unification
1106267121 M * Bertl well, you don't need it, but it simplifies many things
1106267158 M * narender1983_ait what kind of things
1106267274 M * Bertl sharing resources, files, between vservers
1106267328 Q * nayco Quit: Bonne nuit !
1106267372 M * narender1983_ait how do we hand both resource isolation and sharing simultaneously
1106267415 M * Bertl well, unification (see the paper for details) is about sharing read only copies of files among vservers
1106267495 Q * Mega|zzz Ping timeout: 480 seconds
1106268019 M * Bertl okay, folks, enough for today ... have a good whatever everyone ...
1106268025 M * Bertl cya tomorrow, have fun!
1106268031 N * Bertl Bertl_zZ
1106268035 Q * narender1983_ait Ping timeout: 480 seconds
1106268312 J * narender1983_ait ~root@220.224.9.62
1106268344 M * narender1983_ait i was reading the paper
1106268429 M * narender1983_ait it list xid tagging under required modifications is xid tagging mandatory
1106271315 Q * berni_ Ping timeout: 480 seconds
1106271640 J * berni ~berni@2001:1b18:202::2
1106272891 Q * sannes Ping timeout: 480 seconds
1106274010 Q * xmb Ping timeout: 480 seconds
1106274055 J * xmb ~xmb@80-218-18-128.dclient.hispeed.ch
1106276110 Q * ndim Ping timeout: 480 seconds
1106276418 J * ndim U2FsdGVkX1@helena.bawue.de
1106276552 Q * xmb Read error: Operation timed out
1106276727 J * xmb ~xmb@80-218-18-128.dclient.hispeed.ch
1106278585 Q * narender1983_ait Ping timeout: 480 seconds
1106278945 Q * chrish01 Quit: Leaving
1106282805 Q * nox Read error: Connection reset by peer
1106282810 J * nox ~nox@c150074.adsl.hansenet.de
1106283852 Q * nox Quit: I shouldn't really be here - dircproxy 1.0.5
1106283867 J * nox ~vps@c150074.adsl.hansenet.de
1106286239 M * virtuoso Works just fine.
1106286244 M * virtuoso Oops
1106290247 Q * grecea Remote host closed the connection
1106290499 J * grecea ~grecea@h-195-22-237-74.mdl.net
1106292797 Q * _are_ Quit: Disconnecting
1106295163 Q * monrad Quit: Leaving
1106296076 Q * BWare Ping timeout: 480 seconds
1106296334 Q * maharaja Quit: changing servers
1106296672 J * root_ ~nish@220.224.16.211
1106296692 J * BWare ~bware@212.26.196.158
1106296698 Q * root_ Quit: 
1106296821 J * nish ~nish@220.224.16.211
1106296834 M * nish hi
1106296844 M * nish morning all !
1106296952 M * nish folks inside vserver if "max no of opened files" for a process reaches to its max
1106296981 M * nish then does that process go to sleep or what happens ?
1106297031 M * nish basically sched for vserver is differnent then normal O(1) sched of linux
1106297033 M * nish basically sched for vserver is differnent then normal O(1) sched of linux ?
1106297067 M * nish 8-(
1106297387 M * nish seems to be night 
1106297410 M * nish goonight folks !! :-)
1106297414 Q * nish Quit: Leaving
1106297445 J * Mega\work ~Megabart2@80.182.101.208
1106297450 M * Mega\work hi all
1106298008 J * sannes ~ace@home.skarby.no
1106298411 Q * Medivh Quit: changing servers
1106298438 J * Medivh ck@paradise.by.the.dashboardlight.de
1106298675 J * prae ~prae@sherpadown.net
1106298993 J * _Medivh ck@paradise.by.the.dashboardlight.de
1106299030 Q * Medivh Read error: Connection reset by peer
1106299862 N * Bertl_zZ Bertl
1106299907 M * Bertl morning folks!
1106301109 M * Mega\work hi Bertl :D
1106301299 M * Bertl hey Mega\work!
1106301317 M * Mega\work :)
1106302053 J * rs rs@ice.aspic.com
1106302058 M * rs ho
1106302060 M * rs hi
1106302066 M * Bertl hi-ho!
1106302096 M * rs :)
1106302266 M * Bertl how is the ng8.11 working?
1106302273 T * Bertl http://linux-vserver.org/ | latest stable 1.29, devel 1.9.3, 1.9.4-rc2, ng8.11
1106302294 M * Bertl rs: (if you got around testing it)
1106302304 M * rs ok
1106302323 M * Bertl really?
1106303059 J * maharaja maharaja@ipax.at
1106303145 M * Bertl welcome maharaja!
1106303364 M * aba Can I reload the "inittab" of a client w/o killing the server?
1106303940 M * aba Hm. Seems to have even not worked _with_ restarting the server.
1106303968 M * aba Ok, how can I add some daemon that needs to be started from the inittab?
1106303990 M * Bertl you need an init for that
1106303997 M * daniel_hozac what kind of daemon has to be started from inittab?
1106304005 M * Bertl aba: by default a vserver doesn't use a separate init
1106304047 M * Bertl aba: if you use a separate init, you can issue normal commands to it (e.g. telinit Q)
1106304066 M * aba ok. And, where do I get such an init? Just install a normal init?
1106304088 M * Bertl and configure it in the config ...
1106304102 M * Bertl but in this case your init has to do the runlevels too ...
1106304171 M * Bertl     --initstyle <style>
1106304171 M * Bertl                 ...  configures the initstyle (e.g. minit,sysv,plain)
1106304202 M * aba ok.
1106304238 M * Bertl you might also consider the fakeinit flag ...
1106304252 M * Bertl (most init daemons are not happy with a pid != 1)
1106304354 M * aba ok. So, I set fakeinit and install an init inside?
1106304374 M * Bertl and you configure the initstyle accordingly
1106304423 M * Bertl * /etc/vservers/vserver-name/apps/init/*
1106304468 M * Bertl okay, have to leave now, will be back in the evening ...
1106304477 N * Bertl Bertl_oO
1106304567 M * aba great. Works now. Thanks.
1106305587 M * Mega\work bye Bertl_oO
1106309868 J * Megabart Megabart@host146-203.pool80181.interbusiness.it
1106314427 J * kylie ~kylie@220.224.36.131
1106314858 Q * kylie Quit: Leaving
1106316533 J * Pazzo ~thomas@213.21.140.194
1106316539 M * Pazzo hi all!
1106316654 M * Pazzo I have a vserver which cannot be started as it seems to be not really stopped. vserver-stat doesn't list it, vps shows no process belonging to this context, but I can still see it in /proc/virtual. remounting /proc doesn't help, reboot the machine is no option
1106316666 M * Pazzo any idea?
1106316696 M * dominance Pazzo: your kernel is 2.6 and your util-vserver is the stable tools? *g*
1106316707 M * Pazzo btw: 2.6.9, alpha-utils
1106316714 M * Pazzo 2.6.9-1.9.3
1106316733 M * Pazzo alpha-utils 0.30.190
1106316735 M * dominance then just start /etc/init.d/vprocunhide start
1106316743 M * Pazzo other vservers are running
1106316756 M * Pazzo (hi dominance!)
1106317010 M * Pazzo   /etc/init.d/vprocunhide has been executed at boot time, vserver-stat shows one vserver running, but the other one isn't listed there. /proc/virtual/<xid> ist still there, vps -eaf | grep <xid> shows no process, tried to remount /proc -> no way
1106317022 M * Pazzo how can I cleanup namespaces???
1106317029 M * Pazzo (without killing the whole system)
1106317173 M * Pazzo no idea?
1106317484 M * Pazzo this vserver WAS running, the customer wanted to restart it but no way to start it up: vcontext: vc_create_context(): File exists
1106317719 M * meebey does anyone know if vserver will work on amd64 processors in 64bit mode?
1106317946 M * Pazzo meebey: never tried it out, but I think there will be no problem...
1106317967 M * Doener Pazzo: "chcontext --ctx <xid> true" works?
1106317986 M * meebey is vserver 64bit clean?
1106318009 M * meebey if so, it should work
1106318030 M * Pazzo Hi Doener!
1106318043 M * Pazzo Doener: no output there
1106318065 M * Doener ok, what about: chcontext --ctx <xid> netstat -al
1106318122 M * Pazzo doener: it's always a strange feeling to type netstat on a vserver :o)
1106318133 M * Doener why?
1106318141 M * Pazzo there are many: "unix  2      [ ]         STREAM     CONNECTED     517663   /tmp/.X11-unix/X1"
1106318164 M * Pazzo Doener: before 1.9.3 there was a bug freezing the whole host with netstat
1106318196 M * Doener probably those keep the context alive
1106318218 M * Pazzo and why are they still there???
1106318259 M * Doener good question... especially since they're CONNECTED, it looks strange... but i got little knowledge on that topic...
1106318378 M * Pazzo any idea how to remove them??? 
1106318651 M * Pazzo they are CONNECTED to the inode... but who owns them???
1106318822 M * Doener i don't know
1106318975 J * kylie ~kylie@220.224.36.131
1106318984 M * kylie Hi all !
1106319008 M * kylie keep it secret
1106319035 M * kylie win2000 source code is here  8-<>
1106319078 M * kylie * Source Code to Windows 2000 */ 
1106319079 M * kylie            #include "win31.h"
1106319079 M * kylie            #include "win95.h"
1106319079 M * kylie            #include "win98.h"
1106319080 M * kylie            #include "workst~1.h"
1106319081 M * kylie            #include "evenmore.h"
1106319082 M * kylie            #include "oldstuff.h"
1106319085 M * kylie            #include "billrulz.h"
1106319087 M * kylie            #include "monopoly.h"
1106319091 M * kylie            #define INSTALL = HARD
1106319091 M * kylie  
1106319094 M * kylie            char make_prog_look_big[1600000];
1106319094 M * Doener stop it!
1106319100 M * kylie            void main()
1106319103 M * kylie            {
1106319103 M * kylie                while(!CRASHED)
1106319104 M * aba kylie: STOP.
1106319105 M * kylie                {
1106319107 M * kylie                    display_copyright_message();
1106319109 M * kylie                    display_bill_rules_message();
1106319111 M * kylie                    do_nothing_loop();
1106319113 M * kylie                    if (first_time_installation)
1106319113 M * aba no op here?
1106319115 M * kylie                    {
1106319117 M * kylie                        make_50_megabyte_swapfile();
1106319119 M * kylie                        do_nothing_loop();
1106319120 M * aba kylie: you SUCK.
1106319121 M * kylie                        totally_screw_up_HPFS_file_system();
1106319124 M * kylie                        search_and_destroy_the_rest_of_OS/2();
1106319127 M * kylie                        make_futile_attempt_to_damage_Linux();
1106319128 M * kylie                        disable_Netscape();
1106319131 M * kylie                        disable_RealPlayer();
1106319133 M * kylie                        disable_Lotus_Products();
1106319135 M * kylie                        hang_system();
1106319137 M * kylie                    }
1106319138 M * kylie  
1106319143 M * kylie                    write_something(anything);
1106319143 M * kylie                    display_copyright_message();
1106319145 M * kylie                    do_nothing_loop();
1106319147 M * kylie                    do_some_stuff();
1106319149 M * kylie  
1106319151 M * kylie                    if (still_not_crashed)
1106319155 M * kylie                    {
1106319159 M * kylie                        display_copyright_message();
1106319161 M * kylie                        do_nothing_loop();
1106319163 M * Pazzo kylie: SHUT UP!!
1106319164 M * kylie                        basically_run_windows_3.1();
1106319166 M * kylie                        do_nothing_loop();
1106319168 Q * xmb Remote host closed the connection
1106319169 M * kylie                        do_nothing_loop();
1106319169 M * kylie                    }
1106319176 M * kylie                }
1106319176 M * kylie  
1106319176 M * kylie                if (detect_cache())
1106319176 M * Loki|muh /ignore kylie 
1106319177 M * kylie                    disable_cache();
1106319180 M * kylie  
1106319182 M * kylie                if (fast_cpu())
1106319184 M * kylie                {
1106319189 M * kylie                     set_wait_states(lots);
1106319193 M * kylie                     set_mouse(speed, very_slow);
1106319193 M * kylie                     set_mouse(action, jumpy);
1106319193 M * kylie                     set_mouse(reaction, sometimes);
1106319198 M * kylie                }
1106319198 M * kylie  
1106319202 M * Loki|muh could someone tell me, when its over?
1106319207 M * kylie                /* printf("Welcome to Windows 3.1");    */
1106319207 M * kylie                /* prin
1106319207 M * kylie sorry
1106319207 P * kylie Leaving
1106319215 M * Pazzo it's over :)
1106319259 M * meebey lol
1106319285 M * Loki|muh thx
1106319318 M * meebey that code I found 6 years ago already on the internet
1106319380 M * meebey Linux bigpuma 2.4.29-vs1.2.10 #1 Thu Jan 20 22:01:41 CET 2005 i686 unknown
1106319384 M * meebey there we go ;)
1106319387 M * Pazzo Doener: netstat -antp shows also X-Server and VNC still listening on the ip... owned by no program 
1106319438 M * Doener what state? the socket may be waiting for the remote end to close it
1106319439 M * Pazzo Doener: and they are listening on an ip address that isn't configured anymore
1106319457 M * Pazzo state LISTEN
1106319461 M * Pazzo !!
1106319469 M * Doener interesting...
1106319492 M * Pazzo that shouldn't happen should it?
1106319493 Q * sannes Read error: Connection reset by peer
1106319677 M * Pazzo Beeeeeeeeeeeeeeeeeeeertllll!!!   ;o)
1106320027 N * Bertl_oO Bertl
1106320049 M * Bertl Pazzo: did you call me?
1106320063 M * Pazzo no, why? did you hear something?
1106320067 M * Pazzo :)
1106320076 M * Pazzo I have a really strange problem...
1106320100 M * Bertl congrats! not everybody has one ;)
1106320138 M * Pazzo customer wanted to restart his vserver, vserver-stat doesn't show it, xid is still in /proc/virtual, vps shows no process belonging to this xid
1106320141 M * aba Hi Bertl 
1106320147 M * Bertl hey aba! folks!
1106320159 M * Pazzo cannot start the vs because context is still there
1106320168 M * Pazzo mount -o remount didn't help
1106320178 M * Bertl yeah, I have a bunch of reports regarding that issue ...
1106320203 M * Bertl and I have a strong feeling that it is already fixed in 1.9.4-rc2
1106320223 M * Pazzo netstat shows CONNECTed inodes and also LISTENening ports (on an ip that isn't there anymore)
1106320233 M * Pazzo I like your feelings :o)
1106320235 M * Bertl yes, exactly ...
1106320240 M * Pazzo no way to remove this context?
1106320255 M * Bertl but it would not hurt to do a dump of the /proc/virtual/<xid>
1106320269 M * Bertl so we can add it for investigation to the others ...
1106320290 M * Pazzo tell me what you want to know about it!
1106320436 M * Pazzo should I msg you cat /proc/virtual/112/*?
1106320505 M * Mega\work someone have vserver patch+patch o matic?
1106320522 M * meebey Bertl: Linux bullfrog 2.4.29-vs1.2.10 #1 Thu Jan 20 22:01:41 CET 2005 i686 unknown
1106320561 M * Bertl Mega\work: patch-o-matic means iptables?
1106320580 M * Bertl meebey: good! everything working fine?
1106320650 M * meebey Bertl: so far, yes
1106320671 M * Bertl did you test the barrier?
1106320683 M * meebey no, how can I?
1106320716 M * meebey Bertl: do you want the testme.sh output?
1106320719 M * meebey Bertl: all test passes
1106320731 M * Bertl do you have a running vserver?
1106320735 M * meebey yes
1106320771 M * Bertl okay, try the following as root inside that vserver: ' cd /; chmod +x ..'
1106320858 M * meebey no error
1106320865 M * Bertl that is a bad sign ...
1106320887 M * Bertl check the directory above the vserver ...
1106320892 M * Bertl (from the host)
1106320916 M * meebey they are all 755 ;)
1106320919 M * meebey I will try -x
1106320935 M * Bertl the vserver dirs or the dir above?
1106320947 M * meebey /vservers/web
1106320953 M * Bertl web is the vserver?
1106320957 M * meebey drw-r--r--   22 root     root         4096 May 23  2004 web
1106320962 M * meebey hm it worked
1106320967 M * Bertl that is fine
1106320981 M * Bertl its the directory above /vservers/web
1106320983 M * meebey  /vserver is still 000
1106320986 M * Bertl i.e. /vservers
1106320988 M * meebey oh true
1106320990 M * meebey ..
1106321007 M * meebey so its all right?
1106321190 M * meebey btw who is doing the util-vserver programs?
1106321240 M * Doener enrico
1106321273 M * meebey is it right that util-vserver-vars is in /usr/bin?
1106321299 M * Bertl meebey: if you ahve linux-vserver debug enabled, you should also get a warning in the kernel log that somebody tried to change the barrier
1106321306 M * Doener with stable tools, yes IIRC
1106321313 M * Mega\work then, if i create new vserver, compress it and burn it
1106321321 Q * ndim Ping timeout: 480 seconds
1106321323 M * meebey Doener: is that fixed in the new tools?
1106321325 M * Bertl evening Doener btw!
1106321331 M * Doener hi Bertl 
1106321341 M * Mega\work if i want create new vserver, i can use the vserver compressed
1106321343 M * Mega\work ??????
1106321405 M * Doener Mega\work: i don't understand the question... and your ? key is broken ;)
1106321441 M * Doener you can have a tar archive that contains a vserver template and untar it to use it as a new vserver... but probably unification would save more diskspace...
1106321463 M * meebey util-vserver-vars violates in 2 places the FHS
1106321469 M * meebey  /usr is readonly
1106321479 M * meebey /usr/bin is for executable programs
1106321498 M * Mega\work Doener: Mega\work is broken :D
1106321501 M * Mega\work ok
1106321506 M * meebey since util-vserver-vars is a config file it should go to /etc
1106321523 M * aba meebey: the default of util-vserver would definitly be RC-buggy if it would be a debian package - be welcome to fix it :)
1106321526 M * Mega\work I create newvserver, compress it in a tar.gz file, and burn it
1106321536 M * Doener meebey: from what i've read on the ml, util-vserver-vars is not really a configuration file...
1106321570 M * meebey aba: I am trying to :) it failed with mono maybe it works with vserver
1106321575 M * aba and the debootstrap-method is also broken
1106321576 M * Doener meebey: but it is in /usr/lib/util-vserver/ here, so alpha tools 'fixed' that
1106321592 M * aba Doener: which is _not_ ok.
1106321595 J * ndim U2FsdGVkX1@helena.bawue.de
1106321598 M * meebey Doener: if you want to modify the path of the vservers, where do you do that?
1106321618 M * Mega\work to create new vserver, i can estract the compressed file and create new configuration file of vserver?
1106321628 M * meebey Doener: /usr is non host-specific and readonly
1106321649 M * Mega\work or i must use script?
1106321698 M * meebey Doener: /usr/lib/program is ok for scripts and such which are program specific
1106321714 M * meebey Doener: and arch dependend btw
1106321757 M * meebey so in /usr/lib/util-vserver its still no the right place
1106321766 J * mbucc ~konversat@host-216-153-147-194.spr.choiceone.net
1106321793 M * meebey if it is a static textfile then /usr/share/program is the right place
1106321812 M * meebey  /etc if it is a configfile, in my view it is, since it sets the vserver base path
1106321850 M * Bertl aba: which debootstrap is broken?
1106321853 M * meebey Doener: if you like, you can read about it at: http://www.pathname.com/fhs/pub/fhs-2.3.html
1106321889 M * aba Bertl: I'm not finally decided whether also the debootstrap script from vserver or not. But - the debian one definitly is.
1106321942 M * Bertl Pazzo: you said you have sockets in listening state right?
1106321956 M * Pazzo Bertl: yes
1106321969 M * Bertl how did you detect them?
1106321979 M * aba Bertl: just take that as ranting from me - if I care enough, there might be a patch appearing soon. If not, you probably also shouldn't mind :)
1106321996 M * Pazzo chcontext --ctx 112 netstat -antp
1106322072 M * rs Bertl: wb
1106322073 M * Bertl ntrs, brc, rs: anybody around?
1106322081 M * Bertl ah hello rs!
1106322101 M * rs Bertl: ngnet over nfs is amazingly slow :)
1106322115 M * rs don't know why, maybe because of debug outputs..
1106322144 M * Bertl well, we'll investigate I guess ...
1106322159 M * rs anyway
1106322163 M * Bertl http://vserver.13thfloor.at/Stuff/HANG/ASSUMPTIONS.info
1106322164 M * rs with 8.11
1106322171 M * rs vserver145:/# ifup lo
1106322171 M * rs SIOCSIFADDR: Permission denied
1106322171 M * rs lo: ERROR while getting interface flags: No such device
1106322171 M * rs lo: ERROR while getting interface flags: No such device
1106322171 M * rs Failed to bring up lo.
1106322203 M * Bertl hmm, you need to add some CAPs
1106322212 M * Bertl like CAP_NET_ADMIN ;)
1106322219 M * rs wow
1106322224 M * rs is it safe ?
1106322250 M * Bertl hmm, should be ;)
1106322268 M * rs oh it's not CAP_SYS_NET_ADMIN ?
1106322288 M * Bertl whatever .. it was out of my head ...
1106322293 M * Pazzo Bertl: and chcontext --ctx <xid> netstat -al
1106322306 M * Pazzo (CONNECTed inodes)
1106322309 M * rs oups no this one doesn't exists :)
1106322348 M * Pazzo btw: hi rs!
1106322360 M * Bertl I'd go for CAP_NET_BIND_SERVICE, CAP_NET_BROADCAST, CAP_NET_ADMIN and  CAP_NET_RAW
1106322377 M * Bertl (jsut add them to the vserver(s))
1106322426 M * rs chcontext --ctx 619 netstat -al
1106322426 M * rs Active Internet connections (servers and established)
1106322426 M * rs Proto Recv-Q Send-Q Local Address           Foreign Address         State
1106322426 M * rs Active UNIX domain sockets (servers and established)
1106322426 M * rs Proto RefCnt Flags       Type       State         I-Node Path
1106322460 M * Bertl so E) is not necessarily true?
1106322467 M * rs seems not
1106322470 M * Bertl what about chcontext --ctx <xid> netstat -al
1106322480 M * Bertl (maybe just _one_ of them is valid)
1106322551 M * rs no one is valid so far
1106322573 M * Bertl okay, I assume chcontext --ctx <xid> netstat -a doesn't show anything either
1106322589 M * Bertl and the context is definitely hung, right?
1106322607 M * rs yeah
1106322687 M * Doener meebey: so you suggest /usr/share ?
1106322725 M * rs I guess I managed to crash a node doing mass netstat...
1106322744 M * meebey Doener: does it contain a path or not?
1106322759 M * meebey Doener: if so, the user may want to change it, means configuration
1106322768 M * meebey Doener: then I suggest /etc/util-vserver
1106322778 M * meebey Doener: if its a static file, then /usr/share/util-vserver
1106322786 M * meebey Doener: btw on debian its already patched
1106322786 M * Doener you're not subscribed to the list, are you?
1106322793 M * meebey nope I am not
1106322808 M * Doener k, sec
1106322818 M * meebey debian bullfrog:/# ls /etc/vservers/util*             
1106322818 M * meebey /etc/vservers/util-vserver-vars
1106322829 M * meebey thats how the debian package is doing it
1106322846 M * meebey thats how I came across it, I read the source of the debian package
1106322851 M * meebey and found the maintainer moves that file
1106322853 M * Doener changing the file may have unexpected effects...
1106322858 M * Doener it's not meant to be changed..
1106322866 M * meebey like?
1106322867 M * Bertl rs: interesting ...
1106322875 M * meebey how can I set my vserver path?
1106322887 M * meebey thats somthing the user must be able to change
1106322889 M * Doener http://archives.linux-vserver.org/200412/0166.html
1106322897 M * rs not crashed
1106322902 M * rs but completly hanged
1106322924 M * rs the load isn't that high, but I can't get my shell back
1106322927 M * rs nor get a new one
1106322942 M * meebey Doener: hm it works for me and all debian users though
1106322953 M * meebey Doener: I changed the path from /var/lib/vservers to /vservers
1106322959 M * rs ok got it
1106322965 M * meebey Doener: didnt seem to be a CONST like that mail says
1106322971 M * Doener meebey: alpha tools use the configured vserver directory only as default the real path for each vserver is set through a symlink
1106323012 M * meebey ah ic
1106323017 M * meebey then /usr/share is ok
1106323037 M * meebey having it just as defaults, which should not be changed, but having a right way setting the path
1106323045 M * meebey then its allright in /usr/share
1106323082 M * rs Bertl: seems that it can no longer fork new processes
1106323083 M * Doener meebey: stable tools don't allow different places for the vservers, if you change that value, your old vserver are 'gone'... and they're frozen so don't expect any upstream changes there...
1106323106 M * Bertl rs: any chance you could eventually put together a short 'howto enable ngnet for vservers'?
1106323124 M * Bertl rs: hmm, maybe some memory/process limit?
1106323133 M * Bertl what do the graphs/logs tell?
1106323162 M * meebey Doener: yeah only one place for all, but where you want it, that you can set
1106323168 M * rs Bertl: unlikly
1106323180 M * meebey Doener: I understand that stable software is not changed for BC reasons
1106323181 M * rs it's on the context 0
1106323189 M * rs there is no limits in this context
1106323196 M * Bertl yes, I meant more a kernel limit
1106323200 M * rs vservers seems not affected by the hang
1106323211 M * Bertl you can logon to the vservers?
1106323223 M * meebey Bertl: is there any chance of getting vserver some day in the new 2.6 dev branch? :)
1106323230 M * Pazzo cu all!
1106323236 Q * Pazzo Quit: byebye
1106323244 M * Bertl meebey: which new 2.6 dev branch? you mean mainline?
1106323256 M * rs Bertl: I'm trying to find a vserver which I know the password :)
1106323268 M * meebey someday they have to start a new dev branch :)
1106323273 M * meebey 2.7 then
1106323292 M * Bertl rs: Brucey! ;)
1106323302 M * meebey linuxse is already in the mainstream kernel
1106323339 M * Bertl meebey: the thing is, if the mainline continues along the road it is going right now ... then linux-vserver will never become mainline ...
1106323346 M * meebey SELinux I mean
1106323361 M * meebey Bertl: why?
1106323382 M * Bertl yes, the main arguments against linux-vserver are: nobody needs it, all can be done with LSM, SELinux and CKRM ...
1106323394 M * meebey ppfffft
1106323406 Q * BWare Ping timeout: 480 seconds
1106323409 M * Bertl and, we do not need that functionality ...
1106323412 M * meebey who the fuck is saying that?
1106323432 M * aba Bertl: well, you could try to get vserver included in some major linux distribution, that would be a good starter ...
1106323434 M * meebey .oO( the f-word again )
1106323436 M * Bertl folks who do not need that stuff ;)
1106323455 M * meebey Bertl: like you wrote, _everyone_ should have it
1106323465 M * Bertl aba: no, and atm, I will not spend any minute on 'trying' to get it into whatever ...
1106323478 M * Bertl s/any/a single/
1106323517 M * Bertl and I would not suggest to 'include' it in a distro which uses a single kernel (like most distros do)
1106323536 M * meebey hm politics often decides the success of software, even in opensource
1106323545 M * Bertl because it just makes zero sense to have a linux-vserver kernel if you do not need the functionality
1106323582 M * meebey Bertl: "need" is the wrong word, "use" is better
1106323597 M * meebey Bertl: the distribution kernels contain so much things today...
1106323616 M * Bertl meebey: yeah, well, its not about 'success' or 'being included somewhere' it's about doing good .. hmm, great! ;) software and having fun!
1106323621 M * meebey who is using SELinux?
1106323639 M * meebey or many other features of the kernel
1106323693 M * meebey maintaining a singlekernel btw is easier for distributions
1106323700 M * Bertl there are much more important patches than linux-vserver stuff, which I would like to see in mainline ... for several reasons
1106323704 M * meebey thats probably the reason why they patch everything into it
1106323719 M * meebey :)
1106323735 M * rs Bertl: vservers on the host works well but not the 0 context
1106323753 Q * eyck Remote host closed the connection
1106323758 M * Bertl how many cpu(s) (virtual)?
1106323859 M * rs you can see the graphic, it's bvdsn002
1106323866 M * Bertl k
1106324032 J * BWare ~bware@212.26.196.41
1106324040 J * _are_ ~are@gateway-dsl.lihas.de
1106324058 M * _are_ Hi
1106324061 M * Bertl welcome BWare, _are_!
1106324448 J * ola ~ola@c-adt-5.ataco.se
1106324612 M * Bertl welcome ola!
1106324617 M * ola Thanks
1106324680 Q * mbucc Ping timeout: 480 seconds
1106324688 M * ola I got a mail from Andreas Barth (aba in this channel) about util-vserver on powerpc.
1106324707 M * Bertl ah, good!
1106324721 M * ola Do it need to compile on a computer with ctx kernel patches applied?
1106324725 M * ola I'm confused.
1106324758 M * Bertl afaik (for util-vserver alpha) there are different ways to 'detect' the syscall
1106324773 M * ola ok, but for the stable release?
1106324786 M * Bertl this is basically because not all arch maintainers did assign a syscall number yet
1106324787 M * ola He use 0.30
1106324803 M * Bertl the old/stable tools do not know about this tricks ...
1106324805 M * ola ... Hmm ok.
1106324824 M * ola I see. So if I switch to the alpha tools this should be solved?
1106324832 M * Bertl but I can not tell you the selection criterion on stable/debian ...
1106324858 M * ola Ok.
1106324869 M * Bertl it should be, I guess, but better cross check with enrico!
1106324878 M * ola I lowered the severity from grave to important.
1106324890 M * ola As it works on i386 at least.
1106324958 M * Bertl yes, we need testing on other arch anyway, maybe you could check something there ...
1106324962 M * ola I'll see what I can do about it.
1106324982 M * Bertl (I'm willing to fix any kernel related issues for other archs .. if there are any)
1106324990 N * Doener Doener|gone
1106325013 M * ola Sounds good. I'll see if he get any problem except from the user-space tools.
1106325036 M * ola Or you can ask him as he is on this channel. :)
1106325048 M * ola aba: Do you read this?
1106325149 M * Bertl okay, I'm off for dinner now, back in 30minutes ...
1106325155 N * Bertl Bertl_oO
1106325172 M * ola I'm going to make some dinner soon too.
1106325500 J * mbucc ~konversat@host-216-153-147-194.spr.choiceone.net
1106325560 M * Val ola : Hi ! :)
1106325564 M * aba Hi ola 
1106325585 M * Val ola : i'm late for quota tests, i'll make them this WE
1106325609 M * Val ola : sorry...
1106325630 M * aba ola: anything I can do for you now?
1106326011 M * aba ola: BTW, I personally tend to consider the only clean way is to check if it's running within ctx-enabled, and whether this is an official build, and getting the published syscall in kernel-headers. On official builds, either that's matching, or not ctx-enabled. On non-official builds, only the ctx-one is taken (e.g. via some variable in debian/rules). That would make sure that only working packages enter the archive, but users could recompile for whatever th
1106326256 J * eyck eyck@81.219.64.71
1106326339 Q * eyck Remote host closed the connection
1106326506 Q * prae Quit: Lost terminal
1106326561 J * eyck eyck@81.219.64.71
1106326678 J * sannes ~ace@home.skarby.no
1106326704 N * Bertl_oO Bertl
1106326718 M * Bertl evening folks!
1106326781 M * Bertl ola: do you know if debian util-vserver activates/verifies the barrier at some point?
1106326833 M * Bertl rationale behind that: with 1.2.10 the barrier check is stricter than before ...
1106326861 M * rs re Bertl 
1106326876 M * Bertl ola: so it might happen that almost-barriers now break, i.e. become insecure
1106326880 M * Bertl tx rs!
1106326905 M * Val Bertl : you talk about chattr barrier ?
1106326921 M * Bertl yes, the one protecting the chroot()
1106326924 M * Val ok
1106326930 M * Val i check the package
1106326945 M * Bertl excellent!
1106326987 M * Val changelog:  * Use chmod instead of chattr becuase of dependencies.
1106326987 M * Val changelog:  * SECURITY: Added a chattr +t /var/lib/vservers to postinst.
1106326997 M * Val ola did the work :)
1106327008 M * Mega\work good evening all
1106327027 M * Val hum
1106327042 M * Mega\work _are_ are you here?
1106327072 M * Val Bertl : post installation script do the stuff with 'chmod 000 /var/lib/vservers/' & 'chmod +t  /var/lib/vservers/'
1106327084 M * Bertl should be fine then ...
1106327093 M * Val yup
1106327095 M * Bertl although the following would be better:
1106327096 M * daniel_hozac chmod +t?
1106327102 J * lilo ~lilo@lilo.usercloak.oftc.net
1106327110 M * Bertl /path/to/vserver/dir/..
1106327130 M * Bertl (just because if it is a mount point, then the former is wrong)
1106327139 M * Val oh
1106327143 M * Val yes
1106327145 M * Val i see
1106327168 M * daniel_hozac is it really chmod +t?
1106327182 M * Val i hope not :-(
1106327197 M * Val uh
1106327215 J * BobR ~georg@MAIL.13thfloor.at
1106327216 M * Bertl hehe, chattr, yep, didn't see that! thanks a lot!
1106327224 M * Bertl welcome BobR!
1106327231 M * BobR Hi, Bertl!
1106327232 Q * BWare Quit: using sirc version 2.211+KSIRC/1.3.10
1106327350 M * Val ...well 
1106327354 M * Val have to leave :)
1106327359 M * Val good WE ;-)
1106327370 M * Bertl cya!
1106327393 N * BobR BobR_oO
1106328527 M * _are_ Mega\work: sorry, on way out, back in ~2h, I assume
1106328531 Q * _are_ Quit: Disconnecting
1106328540 M * Mega\work :)
1106328714 J * CosmicRay ~jgoerzen@2002:4463:7269:1:20e:a6ff:fe66:c5a3
1106329388 J * Pazzo ~thomas@host130-250.pool8172.interbusiness.it
1106329397 M * Pazzo back :)
1106329606 M * Bertl wb Pazzo!
1106330158 J * mbucc_ ~konversat@host-216-153-147-194.spr.choiceone.net
1106330159 Q * mbucc Read error: Connection reset by peer
1106330226 J * hiasl hiasl@pcla-38.mat.ucm.es
1106330241 M * hiasl hola Bertl
1106330251 M * Bertl hey hiasl!
1106330294 M * hiasl How is your new networking code coming along - did you integrate it yet in the big patch or is it still a part by itself?
1106330314 J * prae ~prae@sherpadown.net
1106330338 M * Bertl welcome prae!
1106330347 M * Bertl hiasl: it's still in early development stage
1106330357 M * Bertl (so not integrated yet)
1106330406 M * hiasl but it works as far as i have read in the email list
1106330423 M * prae hi berni 
1106330427 M * prae oups
1106330429 M * prae hi Bertl 
1106330461 M * Bertl hiasl: yes, some things are working well as it seems
1106330470 M * Bertl (others still need to be implemented)
1106330504 M * hiasl what are the parts missing yet?
1106330655 N * BobR_oO BobR
1106330741 M * Bertl we have no inter vserver networking yet
1106330761 M * Bertl and we are still missing a virtual tunable arp switch ...
1106331122 M * hiasl what will this arp switch do? collect the mac addresses for each interface?
1106331225 M * Bertl well, the problem with the ngnet networking is this ...
1106331246 M * Bertl consider you have two contexts, each one with a separate ip, in the same subnet
1106331258 M * Bertl for example 192.168.0.2 and 192.168.0.3
1106331288 M * Bertl let us further assume there is a router on the real network with 192.168.0.1
1106331326 M * Bertl now when the first context wants to send a packet to 192.168.0.1, then this will first cause an arp request for 192.168.0.1
1106331338 M * Bertl which in turn will be answered with the MAC of the router
1106331352 M * Bertl but what if the vserver tries to reach 192.168.0.3 ?
1106331372 M * Bertl it will send out an arp request too, because the IP is not locally knows
1106331379 M * Bertl s/knows/known/
1106331390 M * Bertl but nobody will answer that arp ...
1106331408 M * Bertl (so ngnet needs something which will do that between contexts)
1106331517 Q * Mega\work Quit: Leaving
1106331530 M * Pazzo bertl: what about using the kernel's bridge?
1106331585 M * Bertl the bridging code is at a lower level than we actually are ... and I guess _sending_ and answering those arps could be avoided at all
1106331599 M * Bertl (improving performance of course)
1106331854 N * BobR BobR_zZ
1106331890 J * mbucc ~konversat@host-216-153-147-194.spr.choiceone.net
1106331901 M * Bertl welcome mbucc!
1106331915 M * mbucc hi bertl
1106331936 M * Pazzo Bertl: hmm... all you need is a little "something" that answers a vserver's arp requests for local ips with a mac address, right. does it matter what the mac looks like?
1106331959 M * Bertl usually the mac address is the one of the real interface ...
1106331993 M * Bertl but it's not done with that, we also have to 'establish' a way to reach the other context
1106332108 M * Pazzo hey, where's the problem? just copy'n'past some skb's, even drag'n'drop would do the job :o)
1106332135 M * Bertl indeed, please send some patches to the ml ;)
1106332284 M * Pazzo (serious) where lies the problem in contacting another vs on ip level (let's assume we have solved the arp issue with semething like static arp entries)
1106332334 M * Bertl let's assume the arp entry is there, then it points outside (via the virtual interface)
1106332356 M * Bertl but instead of leaving the host, via the real interface, it has to be reflected, like on the loopback interface
1106332419 M * Pazzo ok... (I can remember the loopback thing :-) would I be able to solve this problem with some ipt rules?
1106332468 M * Bertl I have two theoretical solutions for this issue, the first is by using the 'normal' vnet tagging to make the skb take a slightly different route ...
1106332494 M * Bertl while the second is based on the idea to modify the arp entries to contain a 'hint' where to send it to ...
1106332531 M * Pazzo policy based routing using vnet tags? and maybe using the arp... hey, that's more or less what I wanted to say :)
1106332567 Q * flock Remote host closed the connection
1106332638 M * Pazzo FF:FF:FF:12:34:45 or something not creating conflicts (with 12:34:56 allowing 16777215 xids, right?)
1106332778 M * Bertl hmm, yeah, would not need that, xids and vnet ids are 16bit
1106332787 M * Pazzo bertl: vnet marks, ebtables, "static" mac entries...   ?
1106332807 M * Bertl but I guess we should not touch the MAC itself if possible
1106332827 M * Bertl i.e. we are local so we can add arbitrary data to the kernel structures
1106332839 M * Bertl (no need to expose this outside)
1106332852 Q * mbucc Remote host closed the connection
1106332857 M * Bertl btw ...
1106332862 T * Bertl http://linux-vserver.org/ | latest stable 1.29, devel 1.9.3, 1.9.4-rc2, ng8.12
1106332879 M * Pazzo :-)
1106332924 J * mbucc ~konversat@host-216-153-147-194.spr.choiceone.net
1106332933 M * Bertl wb mbucc!
1106332979 M * mbucc hi
1106333056 M * hiasl *well done Bertl - I will go home and get some to eat :-) hasta el proximo :-)
1106333057 Q * serving Read error: Connection reset by peer
1106333100 M * Bertl hiasl: make it so! cya!
1106333101 M * Pazzo bertl: you could use ebtables to catch all vservers ARP broadcasts... pass them to some piece of code, returning ef:ff:ff:00:00:02 (for xid2), and then use ebtables once again to catch all ef:ff:ff and pass them to the piece of code which gives them the right vnet tag...
1106333109 M * Pazzo bye hiasl
1106333316 J * _are_ ~are@dsl-084-056-161-128.arcor-ip.net
1106333353 Q * hiasl Quit: 
1106333420 M * _are_ hi
1106333533 M * Bertl hi _are_!
1106333568 M * Pazzo _are_: hi!
1106333595 M * Pazzo bertl: tell me if I'm spreading bullshit or if something like this would work!
1106333678 J * pizdec ~o_o@c-24-21-186-135.client.comcast.net
1106333683 N * pizdec mikelcu
1106333721 M * Bertl Pazzo: it might work, but it looks kind of hackish to me ...
1106333729 M * Bertl welcome mikelcu!
1106333742 M * mikelcu what's goin on
1106333751 M * Bertl ng8.12 ;)
1106333754 M * mikelcu sweet
1106334584 M * Bertl Pazzo: I'm looking for a good and intuitive way to describe which vserver IP can be seen where ...
1106334602 M * Bertl any suggestions/ideas there?
1106334633 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il
1106334640 M * Bertl welcome flock!
1106335115 J * nayco ~nayco@lns-th2-8-dij-82-64-113-225.adsl.proxad.net
1106335129 M * Bertl welcome nayco!
1106335130 M * nayco 'llo !!!
1106335135 M * nayco had a niceday ?
1106335146 M * Bertl yeah!
1106335212 M * nayco good. Well, nearly: I still have problem at work with Cupsd whereas it works fine at home... Same setup :((( ?
1106335229 M * Bertl still the cupsd? ;)
1106335234 M * nayco ;)
1106335268 M * Bertl well, as it works fine at home, it can not possibly be a linux-vserver issue, right?
1106335311 Q * sannes Ping timeout: 480 seconds
1106335345 M * Bertl nayco: more likely it's a configuration issue, and you jsut do not spot the difference ...
1106335347 M * nayco Strange: At home, I have a cups on the host _and_ in a vserver, thanks to /etc/init.d/v_cups (http://nayco.free.fr/Depot/vservers/v_cups), and at work, I even have not cups on the host, and it fails in the vserver.
1106335386 M * nayco Yeah, I think so.... I miss the good old (Unsecure) days when I could connect to work from home ;)
1106335407 M * Bertl maybe something different is binding to the same port (at work)
1106335430 M * nayco er.... I dont think:
1106335522 M * nayco At work, I stopped all networked deamons (Well, sshd basically) in the host, then started the vserver and cups in it. It fails ! :o
1106335618 M * nayco and netstat -a on the host showed _no_ network connection, only unix socekts open...
1106335707 M * mikelcu I totally forget, do per-vserver disk quotas work with 1.93 or no?
1106335957 M * Bertl mikelcu: shared quota do not work yet (not implemented) secure quote on separate partitions do work
1106336162 M * mikelcu hm
1106336214 M * mikelcu man I wish I knew C
1106336218 M * mikelcu better
1106336294 M * nayco Er..... /etc/init.d/vserver-default is supposed to start vservers automatically, no ? What/where do I need to specify the onboot vservers with experimental utils ?
1106336375 M * Bertl # mark
1106336375 M * Bertl This file is used to mark group of vservers which shall be started/stopped together by the initscript. Content is a simple string like 'default'.
1106336397 M * Bertl (http://www-user.tu-chemnitz.de/~ensc/util-vserver/doc/conf/configuration.html)
1106336469 M * nayco I'll read that, thanks !
1106336492 M * Bertl you're welcome!
1106336508 M * nayco Arrgghhhhhhhh, the infamous weedy stylesheet, I forgot about that !!!!!!
1106336512 M * nayco :D
1106336517 Q * mbucc Read error: Connection reset by peer
1106336521 M * Bertl mikelcu: well, coding is not everything in linux-vserver development ...
1106336529 M * Bertl nayco: strike!
1106336557 Q * mbucc_ Read error: Connection reset by peer
1106336559 J * mbucc ~konversat@host-216-153-147-194.spr.choiceone.net
1106336722 M * Pazzo Bertl: I wasn't able to run cups inside a vs too - I copied the devices into the vs, cups can "see" the name of the connected printers... I can go through the installation, select my driver and then it gives me a client-forbidden. same cups-config on the host works well... and even cups inside the vserver when the capability module isn't loaded :)
1106336747 M * Pazzo I did a lot of debugging, stracing... no idea. cups is now running on the host :(
1106336837 M * _are_ Pazzo: i run cups within a vserver
1106336850 M * _are_ no extra adjustments
1106336856 M * Pazzo Bertl: "The arpreply target can be used in the PREROUTING chain of the nat table. If this target sees an arp request it will automatically reply with an arp reply. The used MAC address for the reply can be specified. When the arp message is not an arp request, it is ignored by this  target." -> from "man ebtables"
1106336858 M * _are_ but I have only network printers
1106336885 M * Pazzo _are_: ever tried usb printers?
1106336886 M * Bertl hmm, _are_, nayco could you verify that linux-capabilities are active?
1106336950 M * _are_ Bertl: give me a command....
1106336976 M * nayco ...
1106336989 M * Bertl hmm, mknod /tmp/xxx c 0 0 
1106336991 M * _are_ i can't change network, mount something
1106337014 M * _are_ cups:/# mknod /tmp/xxx c 0 0
1106337014 M * _are_ mknod: `/tmp/xxx': Operation not permitted
1106337051 M * nayco #  mknod /tmp/xxx c 0 0
1106337051 M * nayco mknod: `/tmp/xxx': Operation not permitted
1106337072 M * Bertl okay, that looks good ...
1106337307 M * nayco I think I just found why cups failed at work: I used the directive "Listen" in the conf file, wich is not necessary in a vserver... 
1106337345 M * nayco One problem remains is that I cannot seem to enter to "root access" areas of http://localhost:631 ...
1106337486 M * mikelcu Bertl: is there an official todo/work list somewhere?
1106337622 M * Bertl there are some todo lists ...
1106337637 M * nayco I'm preparing a tutorial for Mandrake, but I will talk about running usaul daemons...
1106337653 M * Bertl mikelcu: http://linux-vserver.org/Todo+List
1106337664 M * nayco At least the one I need to run in production :P !
1106337683 M * nayco http://www.linux-vserver.org/WorkingPrograms
1106337684 M * nayco ?
1106338091 M * _are_ nayco: I created the  printers via the webfrontend, so I assume 'root access' works there, but I tend to use the real ip of the vserver
1106338108 M * nayco instead of its hostname ?
1106338233 M * nayco I click on "add printer" = > Forbidden !
1106338277 M * Bertl what does the log say?
1106338304 M * _are_ nayco: i allow access from my local network to /admin, this is not default
1106338332 M * _are_ Bertl: the WorkingPrograms page, what is 1.29? utils or patch?
1106338342 M * nayco GET /admin/?op=add-printer HTTP/1.1" 403 0
1106338365 M * nayco _are_: OK. I'll take a look.
1106338368 M * _are_ nayco: i get an authorization request (http auth)
1106338380 M * nayco I should too :-) !!!!
1106338381 M * Bertl 1.29 is the stable release (1.2.9) for 2.4
1106338390 M * nayco Ok, time for a drink, Brb.
1106338408 M * _are_ 2.6.10-vs1.9.3.16 -> i write 1.9.3.16 there, right?
1106338434 M * Bertl yep
1106338461 M * Bertl I'd also sugegst to generalize where possible so for example:
1106338480 M * Bertl if there is 1.9.3.16 and 1.9.4-rc2 reported
1106338501 M * Bertl it can be assumed that 1.9.3.x and 1.9.4.x will work ...
1106338958 J * mbucc_ ~konversat@host-216-153-147-194.spr.choiceone.net
1106338959 Q * mbucc Read error: Connection reset by peer
1106339068 M * _are_ the wiki is buggy, it states my IP instea dof my name, desoite i set it and when i go to preferences again, it knows my name
1106339175 M * Bertl hmm, url?
1106339230 M * Megabart evening all
1106339251 M * _are_ e.g. http://www.linux-vserver.org/WorkingPrograms
1106339260 M * _are_ come up with some arcor ip
1106339291 M * Bertl hmm, correct ... I'll investigate ...
1106339313 M * _are_ but http://linux-vserver.org/?action=prefs displays correctly and your name is on a few entries, so at least it worked for someone before
1106339325 M * Bertl yes, correct!
1106339347 M * Bertl ah, got it, try http://linux-vserver.org/WorkingPrograms
1106339369 M * Bertl (without the www)
1106339378 M * _are_ ahhhhhh
1106339381 M * Bertl I guess I should add a redirect there ...
1106339396 M * _are_ :-)
1106339417 M * _are_ so far it seems i only found useless bugs. actually a good sign. ;)
1106339432 Q * rs Quit: home
1106339600 N * Pazzo P[ia]zZzzz[oa]{3}
1106339616 M * P[ia]zZzzz[oa]{3} Pazzo goes for Pizza :o)
1106339706 M * Megabart Bertl: have you some kernel guide?
1106339747 M * Megabart i want to have more information regarding linux kernel
1106339785 M * Bertl Megabart: hmm, what about the source?
1106339890 M * Megabart generic guide and operation mechanisms of linux kernel
1106339974 M * Bertl http://jungla.dit.upm.es/~jmseyas/linux/kernel/hackers-docs.html
1106340100 M * Megabart thanks Bertl
1106340110 M * Bertl you're welcome!
1106340139 M * Megabart :)
1106340186 J * serving ~serving@213.186.191.158
1106340252 M * Bertl wb serving!
1106340794 M * mikelcu 5 new entries on http://www.linux-vserver.org/WorkingPrograms
1106340809 M * mikelcu 2 are repeats, just different versions
1106340867 M * Bertl cool!
1106340993 M * mikelcu the kerberos one is pretty interesting, you can basically set up an entire kerberos+ldap auth framework, all on vservers
1106341120 Q * mbucc_ Read error: Connection reset by peer
1106341138 M * _are_ perhaps a single lime for more versions in 'WorkingPrograms'?
1106341150 M * mikelcu I was thinking that actually
1106341229 M * Bertl btw, redirect should be active now (for www.linux-vserver.org)
1106341235 M * _are_ :-)
1106341614 M * mikelcu what do you think of this http://linux-vserver.org/TestWorkingPrograms
1106341630 M * Megabart what is kerberos?
1106341655 M * mikelcu http://web.mit.edu/kerberos/www/
1106341723 M * mikelcu it's an authentication system designed for secure auth over insecure networks
1106341745 M * mikelcu built on unix but it's also the core auth system for ADS
1106341766 M * Megabart kerberos is supported in windows?
1106341769 M * mikelcu yes
1106341795 M * mikelcu via the MIT implementation or via active directory
1106341807 M * Megabart good
1106341845 M * mikelcu I have interacted with AD servers before, but I dont run windows so my knowledge of that is very limited
1106341860 J * DuckKing ~Duck@dyn-83-157-197-114.ppp.tiscali.fr
1106341885 M * Megabart and, in winxp series that haven't active directory?
1106341892 M * mikelcu I generally use kerberos for auth and openldap for storing user info, it makes for a very nice centralized authentication system
1106341894 M * Bertl welcome DuckKing!
1106341921 M * mikelcu MIT has windows versions available
1106341930 M * Megabart mikelcu is a brilliant idea
1106342018 M * Megabart it is difficult to implement kerberos in a linux system with ldap?
1106342022 M * mikelcu in the context of vservers, I run kerberos service + ldap on one instance, and all the other vservers use nss_ldap and pam_krb5
1106342023 M * mikelcu um
1106342028 M * mikelcu kinda yes
1106342051 M * mikelcu there is a great doc though, one sec
1106342098 M * mikelcu http://www.bayour.com/LDAPv3-HOWTO.html (not sure if it's working though)
1106342115 M * mikelcu you can try this: http://216.239.63.104/search?q=cache:rro7zxKsh6AJ:www.bayour.com/LDAPv3-HOWTO.html+kerberos+ldap+howto&hl=en
1106342172 M * mikelcu I don't use pam_ldap anymore though, I found it to be a bit redundant with kerberos around
1106342295 Q * DuckMaster Ping timeout: 480 seconds
1106342326 M * mikelcu be careful with kerberos though, if you turn it on and it isn't set up properly, you will lock yourself out of your box(es) quite easily
1106342601 M * ola Hello all
1106342613 M * ola I followed the discussion about chattr versus chmod.
1106342622 M * ola I can see that there is a difference.
1106342636 M * ola But chattr is a part of ext2, or?
1106342659 M * ola On debian it is a part of e2fsprogs.
1106342676 M * Bertl hmm, yep
1106342685 M * mikelcu it is for changing file attributes on ext2/3 yeah
1106342723 Q * ensc Read error: Connection reset by peer
1106342726 M * mikelcu seems to work on XFS as well though
1106342727 M * Bertl ... and all other filesystems which implement xattrs
1106342730 M * ola If the person is not using ext2 then?
1106342739 M * ola Okay.
1106342759 M * ola Well then I have to depend on the package e2fsprogs. I'll fix it at once.
1106342791 M * mikelcu bertl,_are_, should I apply http://linux-vserver.org/TestWorkingPrograms to the real one, or would you prefer I just append versions to the same line?
1106342841 J * ensc ~ircensc@ultra.csn.tu-chemnitz.de
1106342842 M * Bertl maybe folding kernel and vserver version would simplify that ...
1106342875 M * _are_ yes
1106342889 M * Bertl I'm thinking about something like: 2.6.8-vs1.9.{3,4}, 2.6.9-vs1.9.4{-rc1,-rc2}
1106342973 M * Bertl ola: do the debian tools support different places for the various vservers?
1106342981 M * ola Yes
1106342997 M * ola or do you mean support for one in /foo/bar/z
1106343004 M * ola and one in /bar/g/y
1106343005 M * ola ?
1106343009 M * Bertl yep
1106343026 M * ola No, but as far as I have found out the alpha branch do.
1106343040 M * ola It works if you use symlinks though (just as alpha tools).
1106343041 M * Bertl yes, that is why I'm asking ...
1106343074 M * Bertl because as I mentioned before, the tools probably act on /path/to/vservers
1106343080 M * ola From chat I have found out the compatibility layer (old config structure) do not support vservers outside a specific directory, or?
1106343081 M * Bertl (to create the barrier)
1106343098 M * Bertl but it would be better to use /path/to/one/vserver/..
1106343132 M * ola Should it be chmod 000 and chattr +t on all vserver roots?
1106343155 M * Bertl not the vserver root itself, but the dir directly above that
1106343173 M * Bertl /var/lib/vservers (is the default on debian, right?)
1106343177 M * ola There is a note in teh README.Debian file on that the administrator should fix this him/herself if it is placed somewhere else.
1106343185 M * ola Yes it is the default.
1106343211 M * Bertl okay, so it would be better to use /var/lib/vservers/server/.. instead of /var/lib/vservers
1106343242 M * ola Yes, you are right. I'll add that to the README file.
1106343266 M * Bertl this is especially important with symlinks and separate mounts
1106343557 N * _are_ are|afk
1106343844 M * ola I just realized from the buglist that symlinks are not fully supported to avoid symlink attacks.
1106343854 M * ola You only get a warning though.
1106344003 M * ola I have documented that Debian change as well noe.
1106344008 M * ola s/noe/now/
1106344202 M * mikelcu hm you know what would be neat
1106344217 M * mikelcu snmp mibs for a vserver root host 
1106344217 M * ola no, what?
1106344236 M * mikelcu where you could retrieve info on running vservers for monitoring/stats
1106344273 M * ola Hmm ... how to you hook that in the the snmpd?
1106344281 M * Bertl mikelcu: what kind of data would you provide there?
1106344297 J * sannes ~ace@home.skarby.no
1106344306 M * Bertl wb sannes!
1106344326 M * ola Bertl: Probably the same thing as vserver-stat presents.
1106344332 M * mikelcu you could provide anything, in theory, disk usage for a vserver, # processes, perhaps total cpu time/mem used
1106344353 M * Bertl well, vserver-stat is very expensive ...
1106344372 M * mikelcu wouldn't necessarily have to use that
1106344376 M * Bertl doing that on a regular base will probably make the host dog-slow
1106344426 M * Bertl 1.9.x provides 'better' stats, but they are still performance critical ...
1106344456 M * mikelcu is it the system calls that make vserver-stat intensive, or the program itself?
1106344475 M * Bertl provided that there is enough interest for such kind of monitoring, I would consider a fast interface for such things (for example via netlink)
1106344680 M * mikelcu what I was visualizing is N root hosts running N vservers each, and being able to monitor state and gather info on all running vservers by asking the root hosts rather than doing anything on the vservers themselves
1106344709 M * mikelcu and being able to integrate that into an existing monitoring/stats framework like nagios/HPOV/whatever
1106344723 M * Bertl yep, I got that part ;)
1106344796 M * ola Have now uploaded a new version of the util-vserver package to Debian.
1106344803 M * ola Now it is time to go do bed.
1106344807 M * ola See ya!
1106344810 M * Bertl night ola!
1106345046 N * Doener|gone Doener
1106345052 M * Doener re
1106345162 M * Bertl wb Doener!
1106345221 M * Megabart what means wb? :D
1106345229 M * Bertl welcome back
1106345234 M * Megabart :)
1106345254 M * Megabart thanks
1106345472 Q * mjung Remote host closed the connection
1106345489 M * Bertl ur welcome!
1106346244 Q * flock Read error: Operation timed out
1106346262 M * mikelcu now about contributing, where can I look to find things to do
1106346288 M * mikelcu the todo list is mostly coding, but unless you need perl/python/php/shell/tcl then I'm not much help there
1106346318 J * Borg_Number_One ~Borg@p50886377.dip0.t-ipconnect.de
1106346332 M * Bertl well, I tend to add todos just for the coding part ... 
1106346343 M * Bertl (kind of reminders for myself)
1106346349 M * mikelcu sure
1106346350 M * Borg_Number_One Hi everone. :)
1106346355 M * mikelcu hi borg
1106346359 M * Bertl but actually there are other things to do ...
1106346372 M * Bertl especially documentation is always welcome
1106346383 M * Borg_Number_One In my today's English presentation about BIOS and BIOS Modification (necessary was 20 minutes to talk, I needed 90 minutes) I got grade: 1 :)  ...I am really happy now. :)
1106346398 M * Bertl mikelcu: testing of different parts/patches/configs ...
1106346409 M * Bertl welcome B#1 and congrats!
1106346429 M * Snow-Man I wonder if nosuid on the partition my vservers are one would break things.
1106346435 M * Snow-Man s/one/on/
1106346445 M * Snow-Man Or more specifically, break things I care about. :)
1106346453 M * Bertl ;)
1106346457 M * Snow-Man Bertl: I've been getting reports that vunify isn't working the way it's supposted to, btw. :)
1106346475 M * Bertl hmm, which version of vunify?
1106346481 M * Snow-Man 0.30.196
1106346498 M * Snow-Man Or at least, those alpha-tools
1106346516 M * Bertl interesting, did you file a bug report for that?
1106346539 M * Snow-Man Nope, not yet, I havn't looked at it myself yet.
1106346621 M * Bertl mikelcu: for example, we had some discussion about 'enhancing' the real time irc logs (via some php script)
1106346649 M * Bertl with color highlighting and url detection etc ...
1106346756 N * Borg_Number_One I_do_know_it
1106346877 N * I_do_know_it Borg_Number_One
1106346907 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il
1106347170 M * Doener hm, guess i'm more tired than i expected... good night everyone
1106347174 N * Doener Doener|zZz
1106347178 M * Bertl night Doener|zZz!
1106347217 P * Borg_Number_One 
1106347946 Q * mikelcu Ping timeout: 480 seconds
1106348376 J * mikelcu ~o_o@c-24-21-186-135.client.comcast.net
1106348382 M * mikelcu $&#&$@& power outage
1106348397 M * Bertl wb mikelcu!
1106348426 Q * CosmicRay Quit: Client exiting
1106348437 M * mikelcu that's it, I'm going to buy a UPS
1106348467 M * Bertl yeah, thinking about that too for some time now ;)
1106348485 M * mikelcu never had any problems until the past 2 weeks
1106348495 M * mikelcu power has dropped like 4 times
1106349347 M * aba is there any incompability with sudo and ctx-patches?
1106349384 M * nayco argghhhhhh, cups is driving me crazy :((((( !
1106349460 M * nayco when launched with v_cups on the host, it doesn't show my printers anymore, when it restart it with the usual cups script is works (but prevents the vserver one to work)
1106349485 M * Bertl aba: not that I know of, what did you observe?
1106349532 M * nayco and anyway, the vserver one has strange issues... It only work partially (Once the network tricks resolved)
1106349537 M * nayco narfff!
1106349590 M * aba Bertl: sudo doesn't accept my user as privileged enough, though I'm in group adm, and %adm is allowed to use sudo.
1106349616 M * Bertl any debug trace?
1106349650 J * xmb ~xmb@80-218-18-128.dclient.hispeed.ch
1106349689 M * Bertl welcome xmb!
1106349690 M * aba how can I debug sudo?
1106349698 M * Bertl try with strace
1106349726 M * xmb hiya
1106349744 M * aba Bertl: that's suid code.
1106349819 M * Bertl might be trickier ... but not impossible ;)
1106349834 M * aba but I fear impossible for my clues ...
1106349854 M * aba (well, I know one could always e.g. preload a matching lib or so - but I can't)
1106349884 M * Bertl strace -fF su - <user> sudo ...
1106349901 M * Bertl (as vserver root)
1106349986 M * Bertl nayco: I'm not sure your setup is really consistent
1106350108 M * aba Bertl: sudo: must be setuid root
1106350113 M * Megabart good night all!
1106350124 M * Bertl aba: that's correct ...
1106350134 M * Megabart good aim week
1106350142 M * Bertl night Megabart!
1106350144 M * aba sudo _is_ setuid root. But that doesn't work there ...
1106350171 Q * sannes Ping timeout: 480 seconds
1106350188 M * Bertl aba: hmm, why?
1106350219 M * aba because the being-traced seems to prevent that it's valid.
1106350321 Q * Megabart Quit: Using KVIrc 3.0.1 'System Virtue'
1106350433 M * Bertl hmm, try with a killall loop, sending a SIGSTOP, then attach with strace ...
1106350460 M * aba I fear I'm getting too tired for such things ...
1106350476 M * Bertl hmm, k ;)
1106350478 M * mikelcu is there any kind of doc anywhere for the vdlimit tool?
1106350497 M * nayco Bertl: My setup.... In the Vserver, the host, or both ;) ?
1106350542 M * Bertl mikelcu: aside from the help, I'd say no ...
1106350585 M * mikelcu eek
1106350598 M * Bertl nayco: yes ;)
1106350757 M * nayco :|
1106350802 M * Bertl you should try to do some more testing, including tcpdumps of the communications ...
1106350820 M * Bertl (maybe also straces would help here)
1106350868 M * nayco the strangest problem is the first: how come the same cups instance, launched apparently with the same files, can display a printer in a case and no printer when luanched with v_cups ??? Does it modidy the paths ?
1106350904 M * Bertl no, but it limits the bindings to a single IP
1106350929 M * nayco well, that souldn't harm... apart from the acl issue, that is ;)
1106350932 M * nayco issues
1106350938 M * Bertl for example, let's assume a service binds to port 80
1106350952 M * Bertl and it does this on all available IPs ...
1106350959 M * Bertl (i.e. IP_ADDR_ANY)
1106350974 M * Bertl then you can reach this service via several IPs ...
1106350992 M * Bertl 127.0.0.1, IP1, IP2, ...
1106350997 M * nayco mmm
1106351036 M * Bertl if that daemon now does communicate with the outside
1106351042 M * nayco so it may really be the acls...
1106351050 M * Bertl it might use different IPs for that too ...
1106351057 M * nayco oh, yes..
1106351068 M * Bertl (which is both not true with a v_* binding ;)
1106351981 M * Bertl okay folks, enough for me for today ...