1144108873 M * matta hi Bertl! 1144109049 M * derjohn daniel_hozac, at least it runs in a non vserver env ... 1144109066 M * derjohn daniel_hozac, now i copy it into a guest ... 1144111416 P * anonc adios 1144112177 J * anonc ~anonc@staffnet.internode.com.au 1144112205 M * Bertl wb anonc! 1144112214 M * anonc hola bertl 1144113545 M * mugwump cat /proc/cpuinfo |egrep '^(model name|cpu MHz|cache size)' | sort | uniq -c 4 cache size : 1024 KB 4 cpu MHz : 2200.000 4 model name : Dual Core AMD Opteron(tm) Processor 275 1144113550 M * mugwump darn 1144113579 M * mugwump Linux magnus 2.6.16-vs2.0.2-rc14sunx4200 #1 SMP Sun Mar 26 07:55:33 NZST 2006 x86_64 GNU/Linux 1144113598 M * mugwump right, first thing's first - time a kernel compile! :) 1144113679 M * micah derjohn: I've got a similar package for sarge 1144113692 M * micah deb http://deb.riseup.net/debian unstable main 1144113704 M * Bertl mugwump: why rc14? 1144113722 M * mugwump I suppose that's terribly old by now ;) 1144113735 M * Bertl of corse :) 1144113763 M * mugwump Basically, because the sysadmins here build kernels based on patches I recommend :) 1144113772 M * Bertl ah, okay :) 1144113782 M * mugwump it's been a couple of weeks since they did it, finally getting around to using it 1144113807 A * mugwump hmms at util-vserver 0.30.204-5sarge2 1144113816 A * mugwump shrugs 1144114271 M * Skram how do i fix the reboot problem? 1144114282 M * Bertl which one? 1144114415 M * Skram when i say reboot inside a vps, it just goes down, not back uo 1144114433 M * Bertl what tool version do you use right now? 1144114459 M * Skram util-vserver 1144114478 M * Bertl 0.30.210? 1144114509 M * Skram [0.30.210-r4] 1144114529 M * Bertl okay, and what kernel version? 1144114536 M * Bertl Skram: is that gentoo? 1144114554 M * Skram hercules vservers # uname -a 1144114554 M * Skram Linux hercules 2.6.14-vs2.0.1-gentoo #1 SMP Fri Feb 3 18:04:56 UTC 2006 i686 Intel(R) Xeon(TM) CPU 3.20GHz GNU/Linux 1144114557 M * Skram Bertl: Yes 1144114598 M * Bertl you might want to update and/or ask Hollow or bonbons about that, it should work fine with 2.0.2-rc15 though 1144114617 M * Skram Hollow: around, dude? 1144114634 M * Skram is there anything i should update in the vserver? 1144114657 M * Bertl no, basically this is a mix of settings, kernel support and proper tools 1144114686 M * Bertl IMHO the tools should be fine, but maybe the kernel is too old to support the gentoo reboot stuff 1144114744 M * Skram my kernel is? 1144114843 M * Bertl I said, maybe, I don't know for sure, but you could try with 2.0.2-rc15 or one of the devel kernels 1144115653 Q * softi42 Ping timeout: 480 seconds 1144116265 M * daniel_hozac derjohn: so, what happened? did it work? 1144116358 M * daniel_hozac Skram: no legacy support in the kernel, i guess? 1144116367 M * Skram *shrug* 1144116370 M * Skram i dont know 1144116412 M * daniel_hozac so, run testme or read /proc/virtual/info 1144116433 M * Skram ? 1144116440 M * Skram VCIVersion: 0002:0001 1144116440 M * Skram VCISyscall: 273 1144116440 M * Skram VCIKernel: 03000034 1144116509 M * Skram where is testme.sh anyways? 1144116531 M * daniel_hozac http://vserver.13thfloor.at/Stuff/SCRIPT/testme.sh 1144116557 M * Skram yeah everything is supposedly working 1144116559 M * daniel_hozac (hmm, i should make these URLs macros...) 1144116564 M * Skram ;P 1144116584 M * daniel_hozac that's not the point. without legacy support, the helper will be called sync. util-vserver doesn't work with that at all. 1144116606 M * daniel_hozac so, either upgrade to 2.0.2-rc15, or rebuild your current kernel with legacy support 1144116680 M * Skram # CONFIG_VSERVER_LEGACY is not set 1144116683 M * Skram okat 1144116708 M * Skram can it be a module? 1144116711 M * Skram ill go look 1144116723 M * Skram doesnt look like it 1144117355 M * daniel_hozac Hollow, micah, phreak``, Bertl, everybody: http://daniel.hozac.com/vserver/util-vserver-0.30.210-vlogin.patch i'm quite certain i've got the signal handling correct this time. 1144117373 M * Bertl sounds good :) 1144117432 M * daniel_hozac previous version, with the signal setup after the child, could cause a left-over vcontext and a zombie vcontext if you'd try to execute a non-existant application. 1144117457 M * daniel_hozac the one before that caused infinite loops every once in a while. 1144117480 M * Bertl yes, very funny stuff :) 1144117483 M * micah daniel_hozac: what is vlogin for? also, I notice that the patch adds vlogin.c with a copyright for enrico, but did you write it? 1144117516 M * daniel_hozac well, i rewrote Hollow's patch. 1144117530 M * daniel_hozac vlogin is for getting a terminal on the inside of the vserver. 1144117540 M * daniel_hozac so things like screen and Debian guests work ;) 1144117564 M * micah ah, so what is an example usage? 1144117576 M * micah vserver exec vlogin? 1144117591 M * daniel_hozac vserver ... enter, exec and start will use it by default. 1144117629 M * daniel_hozac (even though i only added it to OPTS_VCONTEXT_ENTER...) 1144117636 M * micah oh that is useful! 1144117673 M * daniel_hozac to use it manually, add --vlogin to your vcontext commands. 1144117697 M * Bertl any thoughts/tests regarding security? did somebody talk to enrico? 1144117709 M * Bertl (just curious) 1144117739 M * daniel_hozac he commented a bit on Hollow's initial patch, IIRC. 1144117795 M * micah daniel_hozac: in doing a patch I get three hunks with offset -27 lines 1144117805 M * daniel_hozac http://archives.linux-vserver.org/200603/0027.html 1144117851 M * daniel_hozac micah: well, i did diff it against my previous tree (which already had some patches applied). 1144117873 M * Bertl daniel_hozac: so I take it that those comments were addressed? 1144117883 M * daniel_hozac they're not an issue at all in my version. 1144117902 M * daniel_hozac i reimplemented it as an option to vcontext, rather than as a separate program. 1144117915 M * daniel_hozac (as it is in vserver-utils and Hollow's initial patch) 1144117963 M * daniel_hozac both processes are fully inside the guest though, so there shouldn't be any issues. 1144117997 M * micah daniel_hozac: did you submit this to enrico's savanna bug tracker? 1144118001 M * daniel_hozac yeah. 1144118043 M * daniel_hozac https://savannah.nongnu.org/patch/?func=detailitem&item_id=4969 1144118067 M * micah might be good to make a patch based off of enrico's release (rather than based off of your tree with patches) for cleanliness sake 1144118119 M * micah <-- trying to figure out where the fuzz comes from 1144118173 M * daniel_hozac there was fuzz? 1144118185 M * daniel_hozac on a clean tree, there were just some offsets. 1144118229 M * daniel_hozac updated in place, it's now against vanilla 0.30.210. 1144118233 M * micah sorry offsets 1144118265 M * micah I'm confused how hunk 6 could have offsets because hunk 6 is the vlogin.c addition (which doesn't previously exist) 1144118277 M * micah ah, I misread 1144118293 M * daniel_hozac my tree had vcontext-uid applied. 1144118391 M * micah ah right, I have that added during build 1144118575 M * micah daniel_hozac: I'm adding your patch to this build of util-vserver for debian 1144118977 M * daniel_hozac cool. 1144119089 M * daniel_hozac hmm, it seems to not have fixed the issues it should've fixed... investigating... 1144119314 M * micah hmm, if I apply your patch and then try to build I get some compile time errors: 1144119317 M * micah src/vcontext.o: In function `doit':src/vcontext.c:331: undefined reference to `do_vlogin' 1144119347 M * daniel_hozac ah, yes, you need to rerun automake. 1144119374 M * Bertl yes, that was the same I got last time I tried 1144119390 M * Bertl could you add the 'required' changes to the patch? 1144119562 P * matta 1144119566 J * matta ~matta@71.224.125.126 1144119689 M * daniel_hozac http://daniel.hozac.com/vserver/util-vserver-0.30.210-vlogin.patch 1144119912 A * micah gives it a try 1144119961 M * daniel_hozac i had completely forgotten about that, i've been rerunning aclocal and automake for so long. 1144120004 M * micah that compiles a lot better :) 1144120099 M * daniel_hozac humm. it works as expected when i strace it, but won't die if i don't. 1144120529 J * softi42 ~softi@p549D502F.dip.t-dialin.net 1144120544 M * Bertl wb softi42! 1144120569 M * daniel_hozac can anyone spot the race? (or whatever the problem is?) 1144120570 M * Bertl daniel_hozac: so not 100% perfect yet? 1144120596 M * daniel_hozac not quite. 1144121217 M * daniel_hozac i don't understand what's happening, and as it all works when i use strace, i'm not even sure where to look... 1144121396 M * daniel_hozac it's almost as if the SIGCHLD never reaches the parent process. 1144121422 M * Bertl well, strace changes the behaviour significantly 1144121428 M * Bertl (at least inside the kernel) 1144121438 M * matta Bertl 1144121442 M * matta i know i don't talk much 1144121442 M * daniel_hozac right, well, that makes it really hard to debug ;) 1144121449 M * matta but how is ngnet coming along? 1144121471 M * Bertl theory is done, implementation still on hold ... 1144121573 J * gerrit ~gerrit@c-67-160-146-170.hsd1.or.comcast.net 1144121579 M * Bertl welcome gerrit! 1144121631 M * gerrit Hi Bertl 1144121802 M * daniel_hozac oh, of course, SIGCHLD needs to be set before fork... 1144121905 M * daniel_hozac i blame Hollow for that one :) 1144122149 M * matta Bertl: and quota inside vservers is working? 1144122179 M * Bertl yes, but not on a shared partition (though that only requires minor changes) 1144122195 M * Bertl but there was no real demand for that yet 1144122206 M * matta ok\ 1144122341 J * matt1 ~matta@71.224.125.126 1144122389 Q * Loki|muh Remote host closed the connection 1144122394 J * Loki|muh loki@satanix.de 1144122423 M * Bertl wb Loki|muh! 1144122511 Q * mkhl Ping timeout: 480 seconds 1144122771 Q * matta Ping timeout: 480 seconds 1144123408 J * matta ~matta@71.224.125.126 1144123592 M * daniel_hozac Hollow, micah, phreak``, Bertl: sorry for bothering you all... i seem to find new issues all the time. 1144123614 M * Bertl np 1144123847 Q * matt1 Ping timeout: 480 seconds 1144123876 M * Bertl daniel_hozac: I think it' 1144123889 M * Bertl +s a good thing to work on the vlogin 1144123905 M * Bertl it's one of the things folks come up very often with 1144123916 M * Skram vlogin? sounds userful 1144123923 M * Skram what is it supposedly going to do? 1144123972 M * Bertl make vserver .. enter more like login 1144123999 M * Skram cool cool. 1144124034 Q * jake- Read error: Connection reset by peer 1144126896 M * daniel_hozac http://daniel.hozac.com/vserver/util-vserver-0.30.210-vlogin.patch i haven't found any new issues for an entire hour, and i seem to have worked out all the issues i know about. 1144127179 M * Bertl excellent .. you should start versioning that one :) 1144127226 M * daniel_hozac hehe, yeah. 1144128731 Q * weasel jupiter.oftc.net testlink2.oftc.net 1144129088 M * nebuchadnezzar good morning everybody 1144129096 M * nebuchadnezzar lets get work :-) 1144129109 M * Bertl good morning to you :) 1144129130 M * nebuchadnezzar Bertl: do you sleep some time ? 1144129220 M * Bertl yeah, I'm off to bed soon ... 1144129266 M * nebuchadnezzar ok, see you around 18h ;-) 1144129321 M * Hollow morning 1144129359 M * Hollow daniel_hozac: i hate signals :) 1144129620 M * Hollow off to school.. 1144129849 J * weasel weasel@asteria.debian.or.at 1144130409 M * Bertl wb weasel! 1144130802 M * Skram nice domain, or.at 1144130834 M * Skram oh, its an extension sorry 1144133672 J * Smutje ~Smutje@xdsl-87-78-7-189.netcologne.de 1144133863 M * weasel :) 1144134424 M * Bertl okay, off to bed now ... have a nice one everyone ... 1144134430 N * Bertl Bertl_zZ 1144135232 J * ||Cobra|| ~cob@pc-csa01.science.uva.nl 1144136353 J * pagano ~pagano@lappagano.cnaf.infn.it 1144142857 J * mkhl ~mkhl@200-148-40-113.dsl.telesp.net.br 1144142968 J * Smutje_ ~Smutje@xdsl-87-78-42-143.netcologne.de 1144143113 Q * Smutje Ping timeout: 480 seconds 1144143274 Q * ourseven Read error: Connection reset by peer 1144143410 J * ourseven ~ourseven@tor-irc.dnsbl.oftc.net 1144143453 Q * Smutje_ Ping timeout: 480 seconds 1144144068 J * matt1 ~matta@71.224.125.126 1144144431 Q * matta Ping timeout: 480 seconds 1144144475 J * yang ~yang@cpe-213-157-253-172.dynamic.amis.net 1144144532 M * yang I want to create interfaces in the range of 192.168.1.1-50 but i want ip 192.168.1.27 to belong to host, is there a problem if i miss one IP? 1144144716 Q * cryo Ping timeout: 480 seconds 1144145215 M * phreak`` yang: no, there shouldn't be any problem with that 1144145359 J * cryo ~say@psoft.user.matrix.farlep.net 1144145452 M * yang am i able to limit the vserver to 300mb of ram memory only? 1144145635 M * _Gonzo_ yes, you are (don't know how, but you are) 1144145724 J * ourseven_ ~ourseven@tor-irc.dnsbl.oftc.net 1144145836 Q * ourseven Ping timeout: 480 seconds 1144145962 Q * pagano Ping timeout: 480 seconds 1144146373 M * yang another thing, i have created my vserver with option --interface eth0:192.168.1.100/24 instead of real ip, now i changed vserver/interfaces/0/ip to my real IP and server gives me errors... 1144146399 M * yang RTNETLINK answers: File exists 1144146787 J * pagano ~pagano@lappagano.cnaf.infn.it 1144147534 M * yang it works now 1144147848 J * kir_home ~kir@64.119.159.226 1144147978 J * matta ~matta@71.224.125.126 1144148212 Q * matt1 Ping timeout: 480 seconds 1144148497 J * Viper0482 ~Viper0482@p54976988.dip.t-dialin.net 1144148887 J * lilalinux ~plasma@dslb-084-058-200-037.pools.arcor-ip.net 1144149044 J * meandtheshell ~markus@85-124-9-96.dynamic.xdsl-line.inode.at 1144149956 J * Smutje ~Smutje@xdsl-87-78-42-143.netcologne.de 1144150054 J * lilalinux_ ~plasma@h1-gw.of.net-lab.net 1144150481 Q * lilalinux Ping timeout: 480 seconds 1144152671 M * phreak`` Bertl_zZ: if you got a second, would you mind to explain to me what that 0x00000000 addresses (or whatever term that is in C) are ? 1144152706 J * pollux foobar@image4.cpe.fr 1144152723 M * pollux hi 1144152800 M * pollux i am trying the debian packages 2.6.16-1-vserver-686 and got a strange problem: vserver-stat returns: 'chdir(): Permission denied' 1144152852 M * pollux strace show the problem is in: vserver(0xb010000, 0x1, 0, 0x1, 0xffffe400upeek: ptrace(PTRACE_PEEKUSER,19135,44,0): Operation not permitted 1144154531 Q * kir_home Ping timeout: 480 seconds 1144154717 J * _jake- psybnc@murlocs.org 1144154756 M * lilalinux_ 421 Unable to switch capabilities : Operation not permitted 1144154762 M * lilalinux_ I wanted to start the pure-ftpd 1144154923 M * lilalinux_ oh, got it, have to recompile :( 1144155791 Q * cehteh Ping timeout: 480 seconds 1144155846 Q * matta Ping timeout: 480 seconds 1144158722 M * daniel_hozac phreak``: NULL addresses? 1144158833 M * daniel_hozac pollux: showattr -d / 1144158837 J * rusty ~djforrada@86.35.124.246 1144158839 M * rusty hello guys 1144158842 M * daniel_hozac hi 1144158851 M * rusty i have a problem with syslog-ng in vserver guest :) 1144158864 M * rusty Starting system logging: syslog-ngError opening file for reading; error='Operation not permitted (1)', filename='/proc/kmsg' 1144158864 M * rusty Error initializing source driver; source='s_all' 1144158864 M * rusty start failed. 1144158864 M * rusty invoke-rc.d: initscript syslog-ng, action "start" failed. 1144158912 M * daniel_hozac yes, you need to comment that section of the configuration file. 1144159066 M * pollux daniel_hozac: showattr -d / ---Bui- / 1144159067 M * rusty which configuration file? :) 1144159089 M * daniel_hozac pollux: why do you have the barrier set on /? just do setattr --~barrier / 1144159098 M * daniel_hozac rusty: syslog-ng's, obviously. 1144159145 M * pollux daniel_hozac: thanks a lot, it works 1144159160 M * daniel_hozac pollux: did you set the barrier yourself? 1144159166 M * pollux daniel_hozac: I think I'll fill a bugreport on the debian package 1144159175 M * pollux nope, the package installer did it 1144159197 M * daniel_hozac pollux: which util-vserver is that? 1144159213 M * pollux ii util-vserver 0.30.210-2 1144159221 M * pollux (current unstable) 1144159343 M * pollux maybe the vprocunhide script went wrong .. dunno 1144159350 M * pollux i tested on 2 different hosts 1144159351 M * daniel_hozac vprocunhide doesn't do that. 1144159388 M * daniel_hozac did you have any vservers on either host? 1144159399 M * pollux nope, that was a clean install 1144159520 Q * rusty Quit: 1144159598 M * daniel_hozac hmm, yeah, i guess that script change backfired. 1144159600 M * daniel_hozac micah: ping 1144160679 Q * ||Cobra|| Remote host closed the connection 1144162257 M * phreak`` daniel_hozac: hrm, no idea :) thats why I'm asking (fex lib/vserver.h:197:#define VC_VXC_SECURE_MOUNT 0x00010000ull) 1144162276 M * phreak`` (thats from the utils .210) 1144162278 M * daniel_hozac phreak``: well, those are bitmasks. 1144162288 M * daniel_hozac umm, fields, not masks. 1144162308 M * daniel_hozac just another way of saying (1 << X). 1144162377 M * daniel_hozac the hexadecimal are just easier to compare against the output. 1144162408 M * daniel_hozac (like /proc/virtual/X/status:CCaps) 1144162460 M * phreak`` ah, so those (0x00..) are hex ? 1144162509 M * waldi bah, rc15 broke the kernel abi 1144162511 M * daniel_hozac yeah. 1144162520 M * daniel_hozac waldi: hmm, where? 1144162527 M * waldi daniel_hozac: i think vie vx_info 1144162532 M * waldi a/vie/via/ 1144162561 M * waldi i just try to ask genksyms 1144162573 Q * cryo Ping timeout: 480 seconds 1144162609 M * daniel_hozac ah, you mean the internal ABI? yeah. 1144162649 M * waldi yes 1144162765 J * cryo ~say@psoft.user.matrix.farlep.net 1144162895 M * waldi hrm 1144162911 M * waldi i don't find a reference to vx_info in any of the dsymbols 1144162946 M * daniel_hozac it's a type, should it show up? 1144162972 M * waldi yes, genksyms dumps any parsed type 1144163016 M * waldi and there is a change which invalidates most of the symbol versions 1144163086 M * waldi *gna* i should use the correct tree 1144163119 M * daniel_hozac hehe. 1144163248 Q * cryo Ping timeout: 480 seconds 1144163304 M * waldi ah, struct module includes struct task_struct 1144163315 M * waldi and struct task_struct includes struct vx_info 1144163546 J * cryo ~say@psoft.user.matrix.farlep.net 1144163788 M * harry how can i access files inside a mount in a vserver through the host 1144163790 M * harry (for backups) 1144163809 M * waldi vnamespace 1144163915 M * harry what's with that 1144165063 Q * gerrit Ping timeout: 480 seconds 1144165078 P * meandtheshell 1144165317 M * nebuchadnezzar hello 1144166034 J * wenchien ~wenchien@221-169-69-23.adsl.static.seed.net.tw 1144166140 N * Bertl_zZ Bertl 1144166154 M * phedny the interfaces is /etc/vservers/../interfaces, must they have consecutive numbers? 1144166160 M * phedny wb Bertl 1144166168 M * Bertl morning folks! 1144166171 M * Hollow morning Bertl, daniel_hozac! 1144166192 M * Bertl welcome wenchien! 1144166201 M * wenchien Bertl: hi :) 1144166216 M * wenchien problems here.. >_< 1144166254 M * wenchien # vserver dns start 1144166254 M * wenchien access("/proc/uptime"): Permission denied 1144166254 M * wenchien /proc/uptime can not be accessed. Usually, this is caused by 1144166254 M * wenchien procfs-security. Please read the FAQ for more details 1144166254 M * wenchien http://www.linux-vserver.org/index.php?page=Linux-Vserver+FAQ 1144166255 M * wenchien Failed to start vserver 'dns' 1144166306 M * Bertl did you run vprocunhide? 1144166313 M * wenchien vserver works b4 i reboot... 1144166331 M * Bertl again, did you run vprocunhide? 1144166347 M * wenchien Bertl: yes 1144166358 M * Bertl after the reboot? 1144166410 M * wenchien yes 1144166466 M * Bertl okay, output of testme.sh please (i.e. upload it somewhere, e.g. pastebin.com) 1144166512 M * phedny [repeat] the interfaces is /etc/vservers/../interfaces, must they have consecutive numbers? 1144166535 M * wenchien Bertl: http://pastebin.com/640123 1144166548 M * Bertl phedny: no, doesn't matter at all, you can even use names there 1144166572 M * phedny Bertl: okay, then I'll just do opendir() 1144166600 M * wenchien Bertl: http://pastebin.com/640124 1144166608 M * Bertl phedny: iirc, they are created/assigned in _alphabetical_ order 1144166637 M * phedny by the way, soon I'll figure out whether to release my control-panel code open source or not 1144166661 M * Bertl wenchien: hmm, could it be that you have an xid tagged rootfs? or some other patch in the kernel? 1144166686 M * Bertl wenchien: permission denied on grep is not that usual :) 1144166767 M * wenchien Bertl: i don't even know how to use xid.. :P 1144166781 M * Bertl the testme.sh was started on the host, right? 1144166789 M * wenchien Bertl: yes 1144166858 M * wenchien I was using 2.0.2-rc14 and everything is fine, but after i reboot and switch to 2.0.2-rc15, the problem happens... 1144166877 M * wenchien switch back to 2.0.2-rc14, the problem is still there.... 1144166910 M * Bertl could you upload 'cat /proc/slef/mounts' too please? 1144166918 J * bonbons ~bonbons@83.222.39.180 1144166964 M * nebuchadnezzar ok, nothing to patch today ? ;-) 1144166973 M * Bertl wenchien: btw, what do you get when you run vprocunhide right now? 1144166987 M * wenchien Bertl: http://pastebin.com/640142 1144167042 M * Bertl okay, mounts look sane ... so no tagging issues IMHO 1144167055 M * wenchien Bertl: nothing.. 1144167062 M * wenchien # /usr/lib/util-vserver/vprocunhide 1144167064 M * wenchien # 1144167072 M * Bertl okay, now let's try the following: 1144167081 M * Bertl chcontext --xid 45678 egrep 'context|VxID' /proc/self/status 1144167104 M * daniel_hozac wenchien: /etc/init.d/vprocunhide start 1144167118 M * wenchien Bertl: vcontext: execvp("/usr/sbin/vattribute"): Permission denied 1144167153 M * Bertl looks interesting 1144167160 M * wenchien daniel_hozac: I'm using util-vserver 0.30.210-2 from debian sid, there's no more /etc/init.d/vprocunhide 1144167183 M * Bertl even more interesting ... 1144167185 M * daniel_hozac wenchien: umm, i guess that's a bug. 1144167214 M * wenchien Bertl: I'm using 2.0.2-rc15 now, should i switch back to 2.0.2-rc14 ? 1144167239 M * Bertl wenchien: no, I don't think that is relevant 1144167261 M * Bertl wenchien: IMHO there is something weird going on with your filesystems or procfs 1144167271 M * wenchien daniel_hozac: the maintainer integrade initrd scripts into one 1144167274 M * Bertl where I still would opt for procfs 1144167380 M * Bertl wenchien: just to make sure, could you get the 0.30.210 tarball, uninstall the debian package, rebuild with --prefix= (and the other pathes you like to see adjusted) and try the vprocunhide plus testme.sh again? 1144167395 M * daniel_hozac wenchien: just for kicks, what does ls -l /usr/sbin/vattribute, lsxid /usr/sbin/vattribute output? 1144167397 M * micah daniel_hozac: pong 1144167409 M * daniel_hozac micah: see the discussion leading up to the ping. 1144167416 M * micah daniel_hozac: saw it, and have fixed it 1144167423 M * micah daniel_hozac: thanks :) 1144167447 M * wenchien daniel_hozac: # ls -l /usr/sbin/vattribute 1144167447 M * wenchien -rwxr-xr-x 1 root root 17196 2006-04-03 23:18 /usr/sbin/vattribute 1144167454 M * wenchien # lsxid /usr/sbin/vattribute 1144167455 M * wenchien !!ERR!! /usr/sbin/vattribute 1144167465 M * micah daniel_hozac: i'm going to leave the vlogin patch out for now, just to let things settle before uploading to debian 1144167473 J * Loki|muh_ ~loki@213.239.237.100 1144167487 M * wenchien Bertl: ok 1144167543 M * daniel_hozac micah: well, i'm pretty sure it's all worked out now. 1144167560 M * daniel_hozac at least, i can't seem to get it to crash anymore. 1144167675 M * micah daniel_hozac: ok, I've already uploaded a new version without it to fix the reported bugs, I can add it in the next one if you are ready to let a lot of people test it :) 1144167701 M * daniel_hozac micah: Gentoo has used a previous version for a while already ;) 1144167720 M * daniel_hozac as have my Fedora packages. 1144167746 M * derjohn daniel_hozac, I run your (new) patch also inside a guest. works fine. 1144167763 M * daniel_hozac derjohn: great! 1144167783 Q * Loki|muh Ping timeout: 480 seconds 1144167784 M * micah daniel_hozac: where is the latest patch? As it turns out, I haven't actually uploaded this one yet 1144167802 M * daniel_hozac http://daniel.hozac.com/vserver/util-vserver-0.30.210-vlogin.patch?a 1144167814 M * derjohn micah, I didnt know -> you should use the wiki more effectively :) .. Do you have influence on the bind maintainers? 1144167816 M * micah with the ?a on the end? 1144167830 M * daniel_hozac micah: well, doesn't matter. 1144167846 M * derjohn daniel_hozac, I am talkign about the bind stuff ... 1144167846 M * daniel_hozac micah: i just did that to make sure i didn't get a cached version ;) 1144167847 M * micah derjohn: i've replied to the original bind cap bug 1144167853 M * daniel_hozac derjohn: yeah, i figured ;) 1144167861 Q * pagano Quit: Leaving 1144167905 M * wenchien Bertl: oh, btw, # vserver-stat 1144167906 M * wenchien chdir(): Permission denied 1144167925 M * Bertl hmm, hmm ... 1144167933 M * daniel_hozac wenchien: we fixed that already. 1144167938 M * daniel_hozac wenchien: setattr --~barrier / 1144167945 M * Bertl wenchien: that reminds me of old style barrier stuff 1144167953 M * derjohn micah, my bug is better than the original one :) (my is not set wontfix *lol* ) 1144167977 M * Bertl wenchien: maybe something in your boot scripts is getting the barrier completely wrong? 1144167988 M * Bertl micah: any comments on that, maybe? *G* 1144168016 M * derjohn micah, is riseup your private archive? 1144168027 M * wenchien Bertl, daniel_hozac: # setattr --~barrier / then everything works again... @_@ 1144168047 M * micah wenchien: yes that is fixed in the debian -3 package 1144168061 M * micah derjohn: you should merge those bugs if they haven't been merged 1144168066 M * micah derjohn: it is 1144168075 M * daniel_hozac micah, derjohn: they're not really that similar though. 1144168083 M * wenchien micah: @_@ 1144168086 M * Bertl wenchien: okay, case closed, I guess, get the newer debian package and/or do not reboot :) 1144168099 M * micah or just do setattr --~barrier / 1144168109 M * wenchien Bertl, daniel_hozac, micah: thx a lot!! 1144168114 M * derjohn micah, I dont know how to merge then. besides that I was afraid of doing so because of the wontfix status of the old. 1144168176 M * daniel_hozac micah, derjohn: the fix for the vserver one is simple, just get the permitted capabilities and mask the wanted ones. the fix for the "no capabilities support" will be different. 1144168188 M * micah derjohn: the bind9 maintainer is not incredibly responsible about his package 1144168193 M * derjohn daniel_hozac, he's talking about two debian bugs. someoen already reported that bind doesnt start without capa. this was set to wontfix. i opened a new bug instead of marking the old one as pleasefix 1144168207 M * daniel_hozac derjohn: i know, i've read both of them ;) 1144168230 M * derjohn micah, yes, I heared you would make a very fine co-maintainer ;8) 1144168233 M * micah daniel_hozac: how do you mask desired caps? 1144168247 M * daniel_hozac micah: http://daniel.hozac.com/stuff/bind-9.3.2-caps-when-available.patch 1144168260 M * phedny hmm, is it possible to create a UNIX domain socket that crosses a vserver context? 1144168279 M * phedny like the host bind()s on a path in the vserver root and from the vserver a process does a connect() on it 1144168285 M * Bertl phedny: yes, at least in theory, but it's not really easy 1144168294 M * micah daniel_hozac: ah, it requires you to patch bind... same result as compiling with the --no-cap configure option 1144168306 M * daniel_hozac micah: but recompiling with that option removes features. 1144168321 M * daniel_hozac micah: it means you can't run as non-root in a chroot, for instance. 1144168323 M * micah daniel_hozac: i assume you've sent this patch to the debian bug report and to upstream bind? 1144168329 M * phedny Bertl: what would be difficult about that? 1144168343 M * Bertl phedny: both security and isolation issues 1144168362 M * phedny Bertl: but the current kernel patch prevents it or allows it? 1144168368 M * daniel_hozac micah: derjohn sent it to the debian bug report, and i've sent it to ISC (#15929). 1144168384 M * phedny Bertl: not that I need it, so not urgent, but I was just wondering 1144168387 M * Bertl phedny: depends on how and where you do it 1144168428 M * Bertl most 'host' resources can be accessed inside a guest quite fine, and the spectator context sees all of them (in most cases) 1144168445 M * Bertl guest vs guest is isolated and protected 1144168469 M * phedny Bertl: host bind()s on eg. /var/lib/vservers/tmp/some_socket and guest connect()s to /tmp/some_socket 1144168481 M * micah derjohn: which of the two bugs did you send daniel_hozac's patch to? 1144168496 M * micah there are so many bind bugs :P 1144168507 M * micah looks like this one: #360339360339 1144168509 M * micah err, 360339 1144168715 M * micah derjohn: I'll ask the maintainer, I dont really have any influence on him, but might as well boke him 1144168842 M * Hollow Bertl: if i want to check if i have >= 2.1 is this a good way? if (version & 0xFFFF >= 0x0100) ... 1144168939 M * daniel_hozac Hollow: wouldn't version >= 0x00020100 do it? 1144168955 M * Hollow hm.. yeah 1144169198 J * doener ~doener@i5387FDA2.versanet.de 1144169244 M * Bertl welcome doener! 1144169251 Q * wenchien Quit: Leaving 1144169262 J * jesse_ ~wenchien@221-169-69-23.adsl.static.seed.net.tw 1144169273 N * jesse_ wenchien 1144169282 M * Bertl Hollow: yes, the complete > check is the better one 1144169504 M * doener evening! 1144169640 M * ray6 morning Bertl 1144169827 M * Bertl hey ray6! 1144169870 M * Hollow hm, should vx_set_sched fail on 2.0 if set_mask or cpu_id/bucket_id contains values only applicable for 2.1? or just ignore it? 1144169947 M * Bertl the old interface will 'mask' the values 1144169972 M * Hollow so.. just ignore it.. 1144169972 M * Bertl the new interface is not present in 2.0.x yet, IIRC 1144169995 M * Hollow yeah, i check for the interface with vc version 1144170022 M * Hollow but the library contains one struct for both versions 1144170058 M * Hollow so i silently ignore cpu_id and bucket_id on 2.0 1144170117 M * Bertl hmm, you should use an union there, but it should work 1144170241 M * micah daniel_hozac: the debian bind9 maintainer told me "I'll check to see if 2.6 is able to distinguish the two cases yet" 1144170314 M * Hollow Bertl: could you take a look at http://home.xnull.de/work/vserver/libvserver/src/sched.c ? the headers used for compilation are from 2.1 1144170331 M * Hollow it should work with both 2.0 and 2.1 imo 1144170374 J * lafeuil ~lafeuil@4be54-2-82-225-0-42.fbx.proxad.net 1144170468 M * Bertl welcome lafeuil! 1144170493 M * lafeuil Bertl, thanks 1144170540 M * Bertl Hollow: will do, but will take some time, as I'm basically off now (i.e. have to run) 1144170565 M * Bertl so, off for now .. back later! 1144170569 N * Bertl Bertl_oO 1144170570 M * Hollow ok, cu 1144170607 Q * michal` Ping timeout: 480 seconds 1144170778 M * daniel_hozac micah: that's for the "no caps available" case though, right? 1144170926 M * daniel_hozac micah: even if you fix that issue, it still won't work in vservers. 1144170968 J * michal` ~michal@www.rsbac.org 1144171035 Q * restill Quit: Leaving 1144171217 J * lilalinux__ ~plasma@dslb-084-058-200-037.pools.arcor-ip.net 1144171647 Q * lafeuil Quit: Ex-Chat 1144171657 Q * lilalinux_ Ping timeout: 480 seconds 1144172553 J * dearaujo ~dan@pixpat.austin.ibm.com 1144173143 J * ice-t ice-t@81.196.239.134 1144173178 P * ice-t 1144174570 J * Smutje_ ~Smutje@xdsl-87-78-6-165.netcologne.de 1144174664 M * dearaujo anyone able to make a Vserver yum fc4 guest on Mandrake? 1144174678 Q * Smutje Ping timeout: 480 seconds 1144174678 N * Smutje_ Smutje 1144174688 M * daniel_hozac have you installed yum? 1144174702 M * dearaujo yeah installed yum but vserver build complained 1144174713 M * dearaujo then I unstalled and installed the patched yum 1144174749 M * dearaujo then I got a problem saying there is no cElemetTree 1144174753 M * dearaujo during build 1144174774 M * dearaujo btw ,i had to find an rpm for python-abi 1144174788 M * dearaujo not sure where to go from here 1144174825 M * daniel_hozac python-abi is provided by python. 1144174839 M * dearaujo hmm the patched yum complained that it wasnt there 1144174841 M * daniel_hozac you should probably just build it from the source tarball though. 1144174849 M * dearaujo python? 1144174851 M * dearaujo or yum 1144174855 M * daniel_hozac mandrake's python probably doesn't provide it. 1144174857 M * daniel_hozac yum. 1144174860 M * dearaujo k 1144174874 M * dearaujo i grab that and get the patches 1144174874 M * daniel_hozac and you'll need all of the dependencies installed. python-sqlite, python-elementtree, etc. 1144175774 Q * trash Ping timeout: 480 seconds 1144177207 J * gerrit gerrit@dhcp64-134-126-220.sjcc.sjc.wayport.net 1144177368 Q * Viper0482 Quit: bin raus, 1144178701 Q * gerrit Ping timeout: 480 seconds 1144179261 Q * shedi Quit: Leaving 1144179367 J * kir_home ~kir@solution4.masslink.com 1144179641 J * shedi ~siggi@inferno.lhi.is 1144180166 Q * eyck Ping timeout: 480 seconds 1144180407 J * Vanzetti ~nahuel@laf31-2-82-224-249-60.fbx.proxad.net 1144180415 M * Vanzetti hello 1144180571 P * dearaujo 1144180593 M * daniel_hozac hi 1144180624 M * Vanzetti I'm searchings some informations about the stability of the disk quota in the devel version... 1144180635 M * Vanzetti which problem it can provoke? 1144180638 M * Vanzetti any link ? 1144180682 M * daniel_hozac what do you mean? quotas (per user/group) or disk limits (per vserver)? 1144180699 M * Vanzetti per user/group 1144180862 M * daniel_hozac it's the same code as that in stable. 1144180891 M * Vanzetti the 2.1 is not the developement version ? 1144180912 M * daniel_hozac it is. that doesn't mean everything is unstable :) 1144180974 M * Vanzetti ok so it's stable ? 1144180982 M * daniel_hozac yeah. 1144181037 M * Vanzetti http://linux-vserver.org/Release+FAQ <= but there it say that 2.0 ( stable version ) have not the quota support, and 2.1 have, so is it really the same code ? 1144181121 M * daniel_hozac ah, well, vroot is all you need for quota support on different devices. 1144181225 M * Vanzetti are you sure ? something isn't needed at kernel level ? 1144181248 M * daniel_hozac vroot is at the kernel level. 1144181309 M * Vanzetti ok, so we'll test it, thanks for those informations 1144181310 M * Vanzetti :) 1144181346 M * daniel_hozac http://linux-vserver.org/Standard+non-shared+quota in case you haven't found it yet. 1144181410 M * Vanzetti thanks 1144181697 J * eyck eyck@ghost.anime.pl 1144182247 Q * kir_home Ping timeout: 480 seconds 1144183728 Q * bonbons Quit: Leaving 1144184489 Q * lilalinux__ Remote host closed the connection 1144184839 N * _jake- jake- 1144184945 J * aphetadu1 ~aphetadus@quasar.las.ic.unicamp.br 1144184950 P * aphetadu1 1144184992 J * matta ~matta@c-68-32-239-173.hsd1.pa.comcast.net 1144186322 M * Skram anyone know a good spam filter for use with qmail/vpopmail? 1144186324 M * Skram :* 1144187474 M * ray6 are there others than spamassassin? :) 1144187505 M * derjohn bayes-based? bogofilter is pretty popular ... 1144187543 M * derjohn BTW: Hi ray :) long time to hear 1144187573 M * derjohn s/to/no/ 1144189239 M * Skram has anyone used them with qmail? 1144190592 Q * michal` Ping timeout: 480 seconds 1144190907 J * michal` ~michal@www.rsbac.org 1144190934 N * Bertl_oO Bertl 1144190942 M * doener evening Bertl 1144190943 M * Bertl back now ... 1144193373 M * brc BERTL 1144193391 M * brc Any news about disk quota support? :) 1144193395 M * Bertl Vanzetti: sidenote, the quota code in 2.1.1 is slightly different from the one in 2.0.2, but it is a quite old generalization of the mainline quota code, and daniel_hozac is correct that you do _not_ need special code to ahve quota on separate devices 1144193402 M * brc (on the same particion) 1144193427 M * Bertl brc: do you plan to use it in production? 1144193432 M * brc yes 1144193445 M * brc Lot of my users ask for cpanel or other commercial hosting panels. None of them work without quota :( 1144193456 M * Bertl so I take it you would do some testing for that, yes? 1144193475 M * Vanzetti Bertl, But i don't want on a separate device, i want in the vserver virtual root filesystem... 1144193487 M * brc I would prefer to hear from you that it is OK hehe but yes, since my tests won't make me as safe as your answer 1144193497 M * Bertl Vanzetti: there is no virtual root filesystem in Linux-VServer :) 1144193514 M * brc I can do testing 1144193531 M * brc I just wonder why we had the patch for 2.4 and not for 2.6 :P 1144193532 M * Bertl okay, so it looks like we have a few folks interested in the shared quota at all 1144193540 M * Vanzetti Bertl, when you're insine root fs is /dev/hdv1 1144193551 M * Bertl brc: simple, there was _zero_ demand for it ... 1144193559 M * daniel_hozac Vanzetti: that's just mtab. there is no /dev/hdv1... 1144193575 M * Bertl well, there can be, and if, then it's the vroot device 1144193587 M * brc I think that quota is really important. I though that the main purpose for linux-vserver is hosting, if so some of the "big" control panels would help a lot. but they dont work because of the quota :( 1144193611 M * Bertl brc: are you sure that 'quota' is _the_ problem there? 1144193634 M * Vanzetti daniel_hozac, yes it's a fake device 1144193635 M * Bertl brc: I mean, have you tried to enable quota (on a separate partition) and test that? 1144193641 M * brc Yes. They all rely on user quotas 1144193663 M * brc I've installed them on physical servers without any problem. 1144193671 M * brc The panels do work, but they don't measure disk utilization 1144193699 M * Vanzetti brc, it's what i'm searching for 1144193702 M * brc So users or resellers do not know how much they have used and their limitations. 1144193718 M * Bertl Vanzetti: so, are you interested in testing such stuff too? 1144193722 M * brc And their is no limitation , so everyone can start growing their disk 1144193753 M * brc I can do all testing you want 1144193781 M * Bertl JFYI, the per context quota consists of 4 patches, where 3of them are already in vs2.1.1, so it's just the last one which is missing 1144193786 M * Vanzetti Bertl, we just talked with an admin of nfrance ( a french hoster ) and it will put the development version as beta testing for two weeks 1144193856 M * Bertl so I don#t see any big issues with implementing that, it jsut need a good deal of testing ... 1144193939 M * Vanzetti Bertl, we will make it those two week, and I keep you informed 1144193950 M * Vanzetti Well, it's time for me to go to bed, see you soon 1144194072 M * brc i will do good testing 1144194088 M * brc Which would be the tests ? create quotas, make them full, empty quotas, etc ? :) 1144194215 M * Vanzetti and test forcing the write 1144194222 M * Vanzetti with a big load, i think 1144194274 M * brc Vanzetti: Were you trying to get some panel working too ? 1144194298 M * Vanzetti brc, I'm developer of AlternC control panel 1144194312 M * brc url? 1144194314 M * Vanzetti which uses quota ( if enabled ) 1144194319 M * Vanzetti http://www.alternc.org 1144194321 M * brc just found it, everything in french :) 1144194331 M * brc ahh. ok, in pt 1144194333 M * Vanzetti brc, we're french 1144194342 Q * matta Ping timeout: 480 seconds 1144194349 M * brc Is there a demo ? 1144194356 M * Vanzetti http://demo.alternc.org/ 1144194371 M * Vanzetti I just set up it back this afternoon 1144194385 M * brc what's the l/p ? 1144194395 M * Vanzetti admin/admin ( for admin ) 1144194398 M * brc you should make a way of importing/exporting to cpanel. That would get you lot of clients 1144194405 M * Vanzetti virtual/virtual ( for user ) 1144194427 M * daniel_hozac Vanzetti: might want to change the front page, "The Last Version of AlternC" sounds like you're cancelling it. 1144194438 M * Vanzetti brc, but we won't pay for a cpanel ;) 1144194448 M * daniel_hozac that is, the front page of www.alternc.org 1144194461 M * Vanzetti daniel_hozac, ok, i tell it to others 1144194463 M * brc The experience with my clients show that people feel safe with cpanel. If they want they can move to anotehr host (export/import). And they don't want to try other panels.. But if you make it possible to export to CPANEL and import FROM cpanel, everyone would feel safer to use it. if they dont like they can move 1144194511 M * brc Just checked the demo, really nice. you should also work on design with some ICONS and making it really easy to use :) 1144194521 M * Vanzetti yes of course, but we doesn't have any architecture of cpanel 1144194554 M * brc it's just a tgz with public_html , exports from the database, etc 1144194571 M * Vanzetti brc, Yes, but for the moment we are releasing the 1.0 with this design, making it really stable 1144194582 M * brc ok i will check it out when i have time 1144194597 M * brc we have a vps hosting in brazil, we offer our vps's with syscp or vhcs. 1144194599 M * Vanzetti brc, perhaps you can provide me a tgz from cpanel ? 1144194600 M * brc (free control panels) 1144194613 M * Bertl brc, Vanzetti: first step would be to get a test setup where you can start/stop and verify quota accounting and limtis 1144194614 M * brc I am starting with cpanel right now (all users want cpanel) 1144194636 M * Bertl I would suggest based on ext2 first, ext3 and others later 1144194645 M * Vanzetti Bertl, ok i'll transfert this message to the admin 1144194696 M * Bertl brc, Vanzetti: when do you think testing could be done? 1144194709 M * Vanzetti 2 weeks I said 1144194718 M * Vanzetti we will begin the beta test this week 1144194739 M * Bertl okay, so in a few days for two weeks? 1144194741 M * Vanzetti we will test it in beta-prod for 2 weeks 1144194746 M * brc i like ext3 1144194746 M * Vanzetti yes 1144194764 M * Vanzetti Well, see you i'm tired :| 1144194765 M * brc I just need the patch and i can start it tomorrow 1144194771 M * Bertl Vanzetti: good night! 1144194777 M * Vanzetti thanks 1144194778 M * Vanzetti :) 1144194781 M * brc Vanzetti: i will mak a tgz and send you later 1144194783 M * brc Vanzetti: cya 1144194790 M * Vanzetti brc, thanks too 1144194963 M * brc Bertl: does the patch already exist ? 1144195046 M * Bertl no, but it should not take too long as I said 1144195059 M * Bertl so if you like we can start testing tonight :)