1121906019 N * Bertl_zZ Bertl 1121906050 M * Bertl evening folks! (almost expected that ... :) 1121906494 M * brc BERTL! 1121906494 M * brc :) 1121906573 M * Bertl hey brc! 1121906587 M * brc is tehre already something in vshelper to make iptables work inside the vserver ? 1121906625 M * Bertl not that I know of ... 1121906668 M * brc ok 1121906669 M * Bertl but this 'need# might be eliminated with the ngnet stuff sooner or later ... 1121906678 M * brc how is ngnet going? is it stable ? 1121906695 M * brc would it be hard to add iptables to vshelper? I might try that out if you think it is worth it 1121906701 M * Bertl no, not at all ... it's currently kind of dormant ... 1121906765 M * Bertl it's not that hard to get the helper for something like that, you could for example use the sys_reboot right now ... 1121906828 M * brc ok gonna download the sources and check it out 1121906841 M * brc it would be a matter of just adding -s IP or -d IP 1121906842 M * brc isnt it ? 1121906846 M * brc to the command line 1121906891 M * Bertl well, yeah, we had a restart2 (in the helper) it was removed recently 1121906927 M * brc why was it removed ? 1121906930 M * Bertl but it's not that hard to add it back 1121906952 M * Bertl because typically you do not pass 'other' reboot arguments 1121906965 M * Bertl (and it needlessly bloated the patch) 1121906979 M * brc reboot argument? didnt understand 1121906992 M * brc wasn't restart2 just meant to restart the vserver ? 1121907022 M * Bertl you have a kernel source tree at hand? 1121907127 M * brc yes 1121907163 M * Bertl okay, look at kernel/sys.c ~400 (sys_reboot()) 1121907212 M * brc ok hold 1121907234 M * brc ok found 1121907328 M * brc i'm there 1121907330 M * mugwump Hmm, Xen + Vserver is an interesting idea just put forward by a collegue 1121907366 M * mugwump use Xen for the "hard" memory seperation, then Vserver for the admin benefits 1121907402 M * mugwump eg, Xen master, FreeBSD firewall image, Xen images for each security zone 1121907418 M * mugwump then each security zone is a vserver system with multiple servers inside 1121907458 M * Bertl mugwump: yeah, might be interesting .. wanna play with vserver/xen suuport? 1121907476 M * Bertl brc: okay, you see the various functions sys_reboot() does? 1121907490 M * Bertl especially the LINUX_REBOOT_CMD_RESTART2 ? 1121907511 M * brc yes 1121907532 M * Bertl it get an userspace buffer of certain size 1121907544 M * Bertl (256 chars) 1121907558 M * brc ok 1121907569 M * Bertl this or similar could be used without too much changes ... 1121907587 M * Bertl the next better (well, actually _the_ better) approach 1121907611 M * brc dont know if i understood it. it will restart the system AND run a command after it ? 1121907613 M * Bertl would be to use the syscall switch itself (i.e. the vserver syscall) to send iptable commands to the host 1121907647 M * Bertl brc: no, I don't know what it really is for on a real machine .. nobody ueses RESTART2 1121907657 M * brc ok 1121907667 M * Bertl might be to specify some boot arguments or whatever 1121907671 M * brc so i would need to made changes in the kernel ? 1121907684 M * Bertl well, you need to do that anyways ... 1121907684 M * brc and it will call vshelper 1121907686 M * brc is it like that ? 1121907691 M * brc interesting 1121907707 M * Bertl yep, if you check, one of the early checks in sys_reboot() is 1121907711 M * Bertl if (!vx_check(0, VX_ADMIN|VX_WATCH)) 1121907711 M * Bertl return vs_reboot(cmd, arg); 1121907724 M * brc just found it 1121907752 M * Bertl and btw, I 'just' saw a good way to save some space there :) 1121907763 M * brc how ? 1121907770 M * Bertl stack space actually ... 1121907790 M * Bertl if I split that up in do_reboot() and sys_reboot() 1121907808 M * Bertl and just put the buffer into do_reboot() that will save me 256 bytes stack space :) 1121907827 M * brc wont that be free after all ? 1121907863 M * brc that seems to advanced for me. which syscall is the one i should play to do the iptables stuff ? 1121907942 M * Bertl first, to complete my list, the absolutely best approach to pass that to userspace/host would be to intercept the actual iptables commands and send them back to userspace/host 1121907967 M * brc ah k 1121907972 M * Bertl that would not even require a modified iptables command 1121907973 M * brc changes in bash ? 1121908003 M * brc well i am trying to see this as the vserver's user, and i think the better for them is that normal iptables would work 1121908010 M * brc so they can upgrade their system without trouble 1121908048 M * Bertl yes, but AAMOF it's the hardest way to do it from kernel and (host) userspace ... 1121908073 M * Bertl but of course, the most advanced too ... 1121908112 M * Bertl look at an strace of some iptables command ... 1121908142 M * brc ok 1121908156 M * brc at which part ? 1121908172 M * Bertl the part where the info is sent to the kernel ... 1121908316 M * brc getsockopt and those stuff ? 1121908385 M * Bertl yep 1121908416 M * Bertl and the following mmap ... 1121908452 M * Bertl this is the way how userspace communicates with the kernel (in the iptables case) 1121908478 M * brc ok 1121908484 M * brc That seems to complicated 1121908490 M * brc if there was a way to know that iptables is running 1121908493 M * brc hehehe 1121908497 M * brc the binary iptables is runing 1121908500 M * brc and mapping it to some other ting 1121908503 M * brc without replacing the binary 1121908521 M * Bertl hehe, well, that would be a gross hack ... 1121908575 M * Bertl but I guess you see now _why_ it isn't done yet ... 1121908620 M * brc yeah i understand 1121908624 M * Bertl because even if we assume that somehow the arguments to iptables magically pass from the guest to some host helper 1121908646 M * Bertl and (which might be even more complicated) the return codes/values are propably propagated back 1121908694 M * Bertl you still have to make a very smart daemon which actually understands the rules the guest is issuing, to validate and modify (if necessary) them to work in a safe manner for the host 1121908720 M * Bertl otherwise you could simply give the proper permissions and hope for the best :) 1121908803 M * brc heheheh 1121908808 M * brc proper caps you mean ? 1121908814 M * Bertl yep 1121908815 M * brc and how is ngnet going ? 1121908836 M * Bertl well, it's currently broken, and nobody is working on it (yet) 1121908865 M * Bertl and I guess iptables for ngnet will probably take a lot of work ... 1121909108 M * Bertl that might sound discouraging ... but I guess if some folks are really interested, we can arrange something ... 1121909122 M * brc hehehehe 1121909129 M * brc on ngnet or vshelper ? :) 1121909168 M * Bertl depends on a) what solution seems more viable, and b) how much manpower/time/money is available (and in what area) 1121909213 M * Bertl currently most of my time (for linux-vserver) goes straight into the stable release (vs2.0) 1121909939 J * Johnsie ~john@acs-24-154-53-210.zoominternet.net 1121910111 M * Johnsie Hi folks. ;) 1121910129 M * Johnsie The topic is encouraging. 1121910134 M * Bertl welcome Johnsie! 1121910139 M * Johnsie Hey Bertl. 1121910179 M * Johnsie Bertl: You might just be the man with the brains to figure this one out. 1121910208 M * Johnsie I have a vserver host running 2.6.11.9-vs2.0-rc1, Gentoo flavor. 1121910218 M * Bertl hmm, k 1121910225 M * Johnsie I made a vserver tonight and used 'emerge -uD world' to bring it up to grade... 1121910241 M * Johnsie And, it fusses when I reboot it. 1121910241 M * Johnsie http://pastebin.ca/18198 1121910248 M * Johnsie I tried this a few times to be sure. 1121910266 M * Johnsie I made a few vservers up and tested to see... I'm not even sure where to start with that beast. 1121910274 M * Johnsie What am I doing wrong? 1121910277 M * brc 2.0 is from 1.9 ? 1121910307 M * Bertl 2.0 is the stable release (soon to be deployed) of 1.9.x 1121910317 M * Johnsie joybubbles john # uname -a 1121910317 M * Johnsie Linux joybubbles.telephreak.org 2.6.11.9-vs2.0-rc1 #2 SMP Fri May 27 20:10:09 GMT 2005 i686 Intel(R) Xeon(TM) CPU 2.80GHz GenuineIntel GNU/Linux 1121910317 M * Johnsie joybubbles john # 1121910326 M * Bertl brc: we are currently at vs2.0-rc8.1 1121910352 M * Bertl Johnsie: well, let me take an educated guess ... 1121910358 M * Johnsie Mind you, I am not a Linux guru. :/ 1121910365 M * Johnsie I'm sort of at the intermediate level at best. 1121910378 M * Johnsie I do know that it grabbed 2.6.11 headers and what not. 1121910396 M * Bertl try to think about my line of dedcution, and verify that it is valid ... 1121910412 M * Johnsie The host side was updated today and I've restarted it and have older 2005.0 versions of Gentoo running just fine. 1121910415 M * Bertl - you used some gentoo specific 'layout' for the old guest 1121910423 M * Johnsie And some that were ported 2004.3. 1121910445 M * Johnsie Yeah, I started with 2005.0, which was what was suggested, basically. 1121910451 M * Johnsie Should I be doing something different? 1121910457 M * Bertl - this 'old' one probably had some magic from hollow or so ... which uses some is_vserver_guest command 1121910490 M * Bertl - this special check did avoid all the hardware related stuff, if and only if it actually is a vserver guest 1121910511 M * Bertl - now you updated to a new/recent/whatever version 1121910516 M * Johnsie Okay. 1121910531 M * Bertl - and unfortunately this 'overlay?' tool is missing ... 1121910539 M * Bertl # 1121910539 M * Bertl /etc/init.d/checkroot: line 7: is_vserver_guest: command not found 1121910553 M * Johnsie Yeah, it's a little bizarre to me. 1121910561 M * Bertl - this in turn leads to the script failing in several aspects ... 1121910565 M * Johnsie I figured it would have grabbed all applicable dependencies and what not. 1121910579 M * Johnsie Okay. 1121910590 M * Bertl and the bottom line is, you can't do hardware specific stuff inside a guest 1121910607 M * Johnsie Okay. 1121910614 M * Bertl so, as I'm no gentoo guy, and I don't know this overlay/config/whatever thingy 1121910621 M * Johnsie That's alright. 1121910629 M * Johnsie You've given me plenty of insight. 1121910638 M * Bertl I'd look at the gentoo specific pages (from hollow and friends) 1121910639 M * Johnsie Sometimes I just need some direction and a pep talk...haha 1121910664 M * Bertl and I'm pretty sure once you have the 'proper' guest setup, it'll work just fine ... 1121910669 M * Johnsie Okay. 1121910678 M * Johnsie I'll play with it and see what becomes of it. 1121910684 M * Bertl btw, if you want to know what was fixed since rc1, have a look at the updated changelog26 1121910695 M * Johnsie Okay. 1121910709 M * Johnsie By the way, would you suggest I migrate to a newer kernel? 1121910743 M * Bertl a quick & dirty hack just to verify my theory could be to hard-link the /bin/true to /bin/is_vserver_guest :) 1121910755 M * Johnsie Okay. 1121910786 M * Bertl well, we do the rc* thingy to test for the stable release and of course we fix a lot in every rc ... 1121910809 M * Bertl (so it's a good idea to update to rc8.1 but for sure it's not a must) 1121910844 M * Johnsie Okay. 1121911060 M * Johnsie Thank you very much. :) 1121911082 M * Bertl you're welcome! 1121911354 Q * nokoya Ping timeout: 480 seconds 1121912384 M * kevinp Bertl, where do you hide the changelog for 2.0? 1121912419 M * kevinp I'm sure it's hidden in plain sight, but I can't see it 1121912427 M * Bertl it's linked on the main page, indeed ... 1121912457 M * Bertl but the name might be misleading :) 1121912464 M * Bertl http://linux-vserver.org/ChangeLog26 1121912520 M * kevinp yeah, I would think this would be considered up-to-date 1121912527 M * kevinp I was looking on your site anyway 1121912570 M * Bertl well, some guy changed the comment to 'not-so up-to-date' so I had to update it (just to prove him wrong :) 1121912914 M * kevinp lol 1121913083 M * kevinp I updated it to be more accurate 1121913135 M * Bertl good:) 1121913333 M * Bertl ha-ha! I found the super H reference stuff ... 1121913359 M * kevinp I also updated it to say rc8.1 available instead of just rc8, since it started me on the wrong track.. 1121913377 M * Bertl excellent! 1121913488 M * Johnsie Eh? What is the super H reference stuff, if I might ask? 1121913542 M * Bertl processor and instruction manual for the superH series .. sh3/4/5 1121913562 M * Johnsie Oh okay. 1121913570 M * Johnsie Didn't mean to be impolite. 1121913575 M * Johnsie I'm just a curious fellow. 1121913610 M * Bertl np, it is on my list of missing archs ... because I didn't find any docu in the first run 1121913642 J * Doener` ~doener@p54875B4C.dip.t-dialin.net 1121913864 J * nokoya ~young@hi-230-82.tm.net.org.my 1121913873 M * Bertl wb nokoya! 1121914006 M * ddlp howdy 1121914055 M * Bertl hey ddlp! 1121914076 M * ddlp trying to catch up :-) 1121914079 M * ddlp # uname -a 1121914079 M * ddlp Linux debian 2.6.12.3-vs2.0-rc8.1 #1 SMP Wed Jul 20 21:56:33 EDT 2005 i686 GNU/Linux 1121914082 Q * Doener_ Ping timeout: 480 seconds 1121914099 M * Bertl ddlp: and, how is it? 1121914119 A * ddlp has to fix the clock 1121914129 M * ddlp it works so far :-) 1121914727 M * ddlp how do I activate debugging output? 1121914747 M * Bertl at what level? 1121914800 M * ddlp i compiled this one with CONFIG_VSERVER_DEBUG: y 1121914848 M * Bertl okay, so you get entries in /proc/sys/vserver called debu_* 1121914855 M * Bertl debug_* even 1121914866 M * ddlp oky 1121914870 M * Bertl writing non zero values there enables various debug outputs 1121914884 M * Bertl see Documentation/vserver/debug.txt for details 1121914895 M * ddlp cool, i will play with that 1121915290 J * eXplasm explasm@p549F7573.dip.t-dialin.net 1121915299 M * Bertl and even without those settings, some debug output has been enabled 1121915302 Q * mcp Read error: No route to host 1121915326 M * Bertl (for example it will report access to hidden/protected entries/files) 1121915352 J * mcp ~hightower@wolk-project.de 1121915363 M * Bertl wb mcp ! 1121915988 M * ddlp Bertl: i see 1121916508 J * Aiken_ ~james@tooax6-178.dialup.optusnet.com.au 1121916820 M * Bertl morning Aiken_! 1121916836 Q * Aiken Ping timeout: 480 seconds 1121916836 M * Aiken_ good afternoon 1121916854 M * Bertl ah, right, must be afternoon now ... 1121916866 M * Aiken_ 13:34 local 1121917212 M * Aiken_ what have I done wrong? http://pastebin.com/317573 1121917590 M * Bertl hmm, good question ... 1121917601 M * Bertl the first one: # 1121917604 M * Bertl chattr: Inappropriate ioctl for device while reading flags on ... 1121917624 M * Aiken_ nfs maybe 1121917631 M * Bertl looks like your chattr/setattr tools do not support the filesystem, or the filesystem hasn't enabled xattr support 1121917637 M * Aiken_ http_proxy=http://barney:3120 vserver test build --force -m debootstrap -- -d woody -- --arch alpha 1121917652 M * Bertl the second part ... 6. 1121917653 M * Bertl Couldn't work out current architecture 1121917671 M * Bertl looks like a debian or util-vserver issue to me ... 1121917701 M * Aiken_ it needed -- --arch alpha 1121917734 M * Bertl ah, an with that it works? 1121917773 M * Aiken_ I am on my 2nd I: Retrieving Packages 1121917831 M * Aiken_ worked it by setting DEBOOTSTRAP_DIR and runing debootstrap manually 1121918400 M * Aiken_ would the chattr error be the host filesystem or the filesystem the guest is sitting on? 1121918457 M * Bertl most likely the dir right above the guest 1121918609 M * Aiken_ / is ext3 with xattr support compiled in and /vservers is a nfs mount 1121918703 M * Aiken_ nfs is the problem 1121918726 M * Aiken_ found the problem chattr cmd and it works on the host filesystem ok 1121918744 Q * eXplasm Ping timeout: 480 seconds 1121918787 M * Bertl okay .. so your nfs server does not support it ... 1121918790 M * Aiken_ no tail-merging (t) 1121918802 M * Bertl hrm, the tools still use that? 1121918807 M * Bertl what version is this? 1121918816 M * Aiken_ 208 1121918836 M * Bertl in this case, please file a bug report to savannah ... 1121918840 M * Aiken_ vserver-build.functions: chattr -t "$VDIR" 1121918864 M * Bertl hmm, well, okay, it seems to continue anyway, right? 1121918874 M * Aiken_ it continues 1121918879 M * Bertl but it would be smarter to test if it is set first 1121918888 M * Bertl (then remove the old/obsolete flag if found) 1121918893 M * Aiken_ using the kernel nfsd in 2.4.30 1121918935 M * Bertl the only reason for this command is that broken tools got the flags wrong in the older versions and for upgrades from 2.4 to 2.6 1121919071 M * Bertl so yeah, it's fine to do the -t but only if there actually is a 't' to be removed :) 1121919404 M * Bertl hey, just read, reiserfs is going to be cleaned up soon ... 1121919413 M * Bertl (well at least the indentation/formatting :) 1121919859 M * Aiken_ still can not bring myself to move this machine back to resierfs 1121919885 M * Bertl hehe, I can understand that ... 1121919915 J * DaPhreak_ ~phreak@styx.xnull.de 1121919915 Q * DaPhreak Read error: Connection reset by peer 1121919938 M * Aiken_ that episode made me thankfull I had moved from email from pop3 to imap 1121920753 M * Aiken_ is savannah working for you? 1121920759 M * Bertl sec 1121920795 M * Bertl yup, looks fine here 1121921331 Q * Aiken_ Remote host closed the connection 1121921462 J * Aiken ~james@tooax6-178.dialup.optusnet.com.au 1121921480 M * Bertl wb Aiken! 1121921511 M * Aiken either kde or the nvidia driver just died :( 1121921573 M * Hollow morning 1121921596 M * Bertl morning Hollow! 1121921655 M * Bertl Hollow: regarding my monolog with you: http://vserver.13thfloor.at/Experimental/SYSCALL/ 1121921675 M * Hollow monologue? ;) 1121921699 M * Hollow is this gcc4 friendly? 1121921717 M * Bertl don't know, bit it passes --std=c99 :) 1121921734 M * Hollow hm... 1121921735 M * Bertl and, more important, it results in working code :) 1121921747 M * Hollow also with glibc? 1121921765 M * Bertl well, yes, if you set the error handling properly (as usual :) 1121921797 M * Bertl but don't worry, I'll make a special define for that very soon (to simplify it) 1121921813 M * Hollow syscall_new_gnu99.h: line 29-35 there are still named variadic arguments (tail...) 1121921827 M * Bertl take the c99 version, this one is 'gnu' 1121921832 M * Hollow i know gcc complained about this 1121921843 M * Hollow hm, but gnu99 is a superset of c99 1121921863 M * Bertl exactly .. that's why it allows for named variadic macros 1121921900 M * Hollow ah... hm, i thought i used gnu99, but maybe it was c99.. *shrug* 1121921924 M * Bertl the c99 version handles that (but it looks ugly :) 1121921948 M * Hollow tbh, all the syscall stuff looks ugly *g* 1121921958 M * Bertl traitor! 1121921977 M * Bertl :) 1121921987 M * Hollow :P 1121922023 M * Bertl btw, there are patches to update 0.30.208 to this syscall interface 1121922031 M * Hollow it's just that i really don't understand what it's doing *g* 1121922041 M * Bertl http://vserver.13thfloor.at/Experimental/UTIL-VSERVER/ 1121922083 M * Bertl Hollow: simple, issuing a syscall in 12 languages :) 1121922103 M * Hollow for each arch? 1121922119 M * Bertl for most of them .. yes 1121922217 M * Bertl http://vserver.13thfloor.at/Experimental/SYSCALL/syscall_new.txt 1121922231 M * Bertl all archs listed with [0] are supported right now ... 1121922256 M * Hollow quite a few ;) 1121922259 M * Bertl missing are: cris, the entire SuperH range, and mips 1121922281 M * Bertl ah, and m68k ... 1121922315 M * Bertl yep, and every arch has it's own registers and calling conventions 1121922375 M * Hollow yeah, i know the basic in theory, but i just don't know what all this assembler code does.. ;) 1121922399 M * Bertl half of the file is just comments :) 1121922485 M * Hollow and moreover i didn't get the error thing as well... 1121922565 M * Hollow in libvserver-0.2 there is a choice between glibcs syscall() (which works with shared objects) and your syscall implementation but only using static libraries 1121922573 M * Hollow so this is only semi-fixed 1121922591 M * Bertl we'll fix that, for both dietlibc threaded and glibc 1121922627 M * Hollow ok 1121922667 M * Bertl /usr/include/bits/errno.h 1121922679 M * Bertl check the define __set_errno(val) 1121922686 M * Hollow it's just, that often i simply don't know where to look for what as C beginner.. 1121922699 M * Hollow sec.. 1121922710 M * Bertl basically it should work out of the box if you define: 1121922745 M * Bertl __syscall_error(e) do { __set_errno(e); } while(0) 1121922755 M * Bertl #define __syscall_error(e) do { __set_errno(e); } while(0) 1121922783 M * Hollow and why is that so? 1121922792 M * Bertl before you include the syscall_new.h 1121922811 M * Bertl thing is, if you have static code, then errno is a static variable 1121922820 M * Bertl it has a well known location 1121922839 M * Bertl if you switch to dynamic code, you have to 'find' the variable first 1121922853 M * Bertl which is done by __errno_location() 1121922872 M * Hollow find it where? in memory? 1121922876 Q * Johnsie Quit: 1121922883 M * Bertl yeah, basically ... 1121922886 M * Hollow ok 1121922929 M * Hollow and __set_errno gives the location? 1121922945 M * Bertl no, that does get the location, and stores the value there ... 1121922960 M * Bertl #define __set_errno(val) (*__errno_location ()) = (val) 1121922965 M * Hollow ah 1121922987 M * Bertl but it should be provided by your libc includes 1121922996 M * Bertl and it should also work for dietlibc, btw 1121923000 M * Hollow hm.. at least not in bits/errno.h 1121923054 M * Bertl well, here it is ... but I don't know if that is default ... 1121923101 M * Hollow hm, i grepped thru /usr/include, no such define here 1121923131 M * Bertl with -r ? 1121923136 M * Hollow yup 1121923141 M * Hollow grep set_errno -r . 1121923158 M * Bertl grep for __errno_location 1121923170 M * Bertl the set might be 'obsoleted' by now ... 1121923180 M * Hollow this one is in bits/errno.h 1121923208 M * Bertl so using the indirection should be fine then ... 1121923239 M * Hollow indirection? 1121923277 M * Bertl (*__errno_location()) = (val) 1121923296 M * Hollow in __syscall_error? 1121923322 M * Bertl yep .. maybe try with: 1121923342 M * Bertl #define __syscall_error(e) do { (*__errno_location()) = (e); } while(0) 1121923373 M * Bertl as I said, not tested yet, but modulo some defines that should work fine ... 1121923387 M * Bertl make sure to #undefine __syscall_error before ... 1121924152 M * Hollow Bertl: doesn't work 1121924168 M * Bertl doesn't compile, or bugs? 1121924177 M * Hollow still segfault 1121924214 M * Bertl could you use -E to write assembler output for the code in question? 1121924261 M * Hollow for gcc? 1121924281 M * Bertl yep, for a simple test program which segfaults 1121924308 M * Bertl or use objdump -d and upload the output somewhere 1121924419 M * Hollow Bertl: http://home.xnull.de/misc/libvserver_dump.txt 1121924483 M * Bertl which one segfaults? 1121924489 M * Bertl vx_??? 1121924491 M * Hollow in this case vx_create 1121924517 M * Hollow http://phpfi.com/70835 1121924522 M * Hollow first one with your syscall 1121924526 M * Hollow second one with glibc syscall 1121924600 M * Bertl well, I hate to say it, but there is no syscall in that dump? 1121924644 M * Hollow hm.. 1121924685 M * Bertl okay, where is the source? preferable as tar? 1121924766 M * Hollow http://home.xnull.de/misc/libvserver-0.2.1.tar.bz2 1121924804 M * Bertl how to build? 1121924815 M * Hollow ./configure --enable-alt-syscall && make 1121924832 M * Bertl doesn't look like I have a configure yet 1121924844 M * Hollow oh... it's svn checkout... make -f Makefile.cvs 1121924871 M * Bertl automake -afc 1121924871 M * Bertl automake: unrecognized option -- `-afc' 1121924889 M * Hollow wait.. 1121924910 M * Hollow ok, load the tar again 1121924965 M * Bertl k, configure is running ... 1121924984 M * Bertl ls 1121924989 M * Bertl *oops* 1121925546 M * Bertl vcontext is a shell script? 1121925597 M * Hollow Bertl: either make install or use tools/.libs/lt-vcontext 1121925657 M * Hollow the shell script is a wrapper for the LDPATH (libvserver.so is not installed yet) 1121925667 M * Bertl k, tx 1121926921 M * Bertl well, something goes definitely wrong here ... 1121926931 A * Hollow nods 1121926953 M * Bertl look (here the code _before_ it is run ... 1121926964 M * Bertl 0x40014f2e :mov 0xc(%ebp),%ecx 1121926964 M * Bertl 0x40014f31 :mov 0x10(%ebp),%edx 1121926964 M * Bertl 0x40014f34 :mov 0x8(%ebp),%ebx 1121926967 M * Bertl 0x40014f37 :int $0x80 1121926969 M * Bertl .. 1121926977 M * Bertl and in the error case: 1121926984 M * Bertl 0x40014f4b :mov %eax,%edi 1121926984 M * Bertl 0x40014f4d :call 0x40014bec 1121927003 M * Bertl (this is supposed to be the __errno_location() call 1121927032 M * Bertl okay maybe that one ... 1121927033 M * Bertl 0x40014f5e :call 0x40014bfc 1121927041 M * Bertl anyway ... 1121927051 M * Hollow and what's wrong? :) 1121927075 M * Bertl (gdb) disass __errno_location 1121927075 M * Bertl Dump of assembler code for function __errno_location: 1121927075 M * Bertl 0x4000de60 <__errno_location+0>:push %ebp 1121927090 M * Bertl it is simply at a different location :) 1121927095 M * Hollow heh 1121927101 M * Bertl while 1121927118 M * Bertl (gdb) disass 0x40014bec 1121927118 M * Bertl No function contains specified address. 1121927134 M * Bertl which of course leads to a segfault ... 1121927155 M * Hollow so errno_location returns a wrong adress? 1121927172 M * Bertl it seems that this happens at link time ... not sure how ... 1121927178 M * Bertl (or even why) 1121927209 M * Hollow hm, maybe some define for glibc is missing? 1121927629 Q * Hollow Remote host closed the connection 1121927643 J * Hollow ~Hollow@home.xnull.de 1121927661 M * Hollow .. 1121927731 Q * Hunger arion.oftc.net europa.oftc.net 1121927752 J * Hunger Hunger.hu@Hunger.hu 1121928451 Q * mcp Ping timeout: 480 seconds 1121928625 M * Bertl hmm ... 1121928628 M * Bertl U __errno_location@@GLIBC_2.0 1121928632 M * Bertl w __cxa_finalize@@GLIBC_2.1.3 1121928746 M * Hollow i.e.? 1121928818 J * mcp ~hightower@wolk-project.de 1121929354 Q * nokoya Quit: changing servers 1121929369 J * nokoya young@hi-230-82.tm.net.org.my 1121930013 M * Hollow gtg to school, cu later 1121930017 M * Bertl k, cya! 1121930311 J * DaPhreak ~phreak@styx.xnull.de 1121930448 Q * DaPhreak_ Read error: Connection reset by peer 1121931240 M * Bertl okay, off to bed again ... back later ... 1121931246 N * Bertl Bertl_zZ 1121931584 J * prae ~prae@ezoffice.mandriva.com 1121931861 J * bipsen ~secret@pat.progressive.dk 1121931936 M * bipsen Tried to compile 2.6.12.3-vs2.0-rc8 - but the kernel panics during boot (the screen output scrolls so fast, so I can't see exactly where the stack/code trace/breaks are)... has anyone seen this before ? 1121932177 M * Aiken I skipped 8 and went to straight to 8.1. Have not have a problem with it yet 1121932413 M * bipsen Just noticed the 8.1 - trying to build a new kernel with it right now 1121935362 Q * Aiken Ping timeout: 480 seconds 1121937304 J * morfoh ~jeru@mail.oberlinhaus.de 1121937334 M * morfoh hi all 1121937499 J * eXplasm explasm@p549F7573.dip.t-dialin.net 1121937651 J * erwan_taf ~erwan@81.80.43.77 1121937857 P * erwan_taf 1121938836 J * bips ~secret@pat.progressive.dk 1121938836 Q * bipsen Read error: Connection reset by peer 1121939305 N * bips bipsen 1121940261 J * richardw ~richard@wlan-tiwag-231-207.utaonline.at 1121940272 M * richardw hi 1121940571 J * ps ~ps@fw-lan-transit.le1.spacenet.de 1121940575 M * ps hi 1121940667 M * ps i need to run samba inside a vserver. i did not patch samba until now - if i start the samba vserver after few minutes the load is going very high. is that the "oplock" problem? 1121941077 Q * virtuoso Ping timeout: 480 seconds 1121941288 J * virtuoso ~s0t0na@80.253.205.251 1121943225 Q * eXplasm Remote host closed the connection 1121943268 J * eXplasm explasm@p549F7573.dip.t-dialin.net 1121943547 Q * virtuoso Ping timeout: 480 seconds 1121943923 J * virtuoso ~s0t0na@80.253.205.251 1121943961 Q * richardw Read error: Connection reset by peer 1121944241 J * richardw ~richard@wlan-tiwag-231-207.utaonline.at 1121944754 J * chand ~chand@staff.lycos.fr 1121944759 M * chand hey all 1121945030 M * richardw hi chand 1121946997 M * chand hey richardw sup ? 1121947106 M * richardw hmm? 1121947182 M * chand how u doing ? 1121947241 M * richardw very well, thanks. you too? 1121947333 N * BobR BobR_afk 1121947467 M * chand i'm fine .. a little problem with vserver but well :) 1121947515 M * FaUl hi you 1121948350 J * ptonguet ~ptonguet@nat1.cur-archamps.fr 1121948366 J * ptonguet_ ~ptonguet@nat1.cur-archamps.fr 1121948377 M * ptonguet_ hello, does anybody install ltsp on a vserver? 1121948403 P * ptonguet 1121948417 M * brc Good morning 1121950807 Q * Vudumen Ping timeout: 480 seconds 1121951024 J * Vudumen vudumen@perverz.hu 1121951580 Q * morfoh Remote host closed the connection 1121951894 J * hiaslboy ~matthias@pppsz1146.tu-graz.ac.at 1121952030 Q * hiaslboy Quit: 1121952993 J * erwan_taf ~erwan@81.80.43.77 1121953270 P * erwan_taf 1121955142 J * rivy ~rivy@sep.rivy.org 1121955265 M * rivy someone available? 1121955274 M * rivy I have a question 1121955684 M * aba rivy: usually, you just better just ask 1121955725 M * rivy ok, so I have a box on the internet with 1 external IP 1121955738 M * rivy every vserver on it has a private IP 1121955756 M * rivy like 192.168.0.1 etc 1121955824 M * rivy if another box on my switch uses the same private ip addresses, they will get an error 'double ip' because my box replies on arp packets for his private ip 1121955869 Q * ptonguet_ Quit: Leaving 1121955878 M * rivy what's the best way to stop the box from replying to arp broadcast for one of his private ips 1121955913 M * rivy in fact, the private ips cannot be visible outside the box 1121956115 M * aba make a "virtual" interface and bound the ips to that one? 1121956180 M * rivy you mean with tun/tap? 1121956190 M * aba more like dummy. 1121956227 M * rivy hmm, I'll check it out later. But then another thing. 1121956242 M * rivy the vserver share the routing table with the real host 1121956268 M * rivy but the vservers should route their outbound traffic to the dummy 1121956294 M * rivy and the real host to eth0 of course 1121956322 M * rivy so the 2 route table wouldn't be the same 1121956332 M * rivy s/table/tables/ 1121956363 M * rivy BTW: I'm new to vservers so it could be that I'm making big mistakes 1121956413 M * aba rivy: IIRC there is ngnet for that - but don't ask me, I'm not an expert on that 1121956819 M * rivy thanks, now I have at least something to google for 1121956961 Q * bipsen Quit: 1121957227 J * Snow-Man ~sfrost@snowman.net 1121957589 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il 1121958866 Q * wurd Quit: D 1121958914 N * BobR_afk BobR 1121959149 N * BobR BobR_oO 1121959278 J * prae_ ~prae@gut75-1-81-57-27-189.fbx.proxad.net 1121959721 Q * prae Ping timeout: 480 seconds 1121960496 Q * flock Ping timeout: 480 seconds 1121960678 M * FaUl does nmbd need capabilitys? 1121960685 M * FaUl (the samba nmbd) 1121960783 M * Greek0 probably some network caps, so it can bind below 1024 1121960817 M * FaUl i thought maybe it must send broadcast or something 1121960935 M * Greek0 well, just try it. CAP_NET_BIND_SERVICE and CAP_NET_BROADCAST might be what you need. 1121961060 M * Greek0 strace might help you to see which syscall failed exactly, to see why it isn't working 1121961147 M * FaUl yes, i'll try it 1121961551 J * wurd ~kvlt@modemcable181.93-202-24.mc.videotron.ca 1121961770 J * richardw_ ~richard@wlan-tiwag-231-116.utaonline.at 1121961797 Q * prae_ Quit: Execute Order 69 ! 1121962151 Q * rivy Quit: BitchX: treats external genital warts! 1121962212 Q * richardw Ping timeout: 480 seconds 1121962359 N * richardw_ richardw 1121963212 Q * eXplasm Read error: Connection reset by peer 1121963234 J * eXplasm explasm@p549F7573.dip.t-dialin.net 1121963576 N * richardw richardw|na 1121963843 Q * ps Remote host closed the connection 1121964130 Q * wurd Quit: BitchX-1.0c20cvs -- just do it. 1121965227 N * Bertl_zZ Bertl 1121965256 M * Bertl morning folks! 1121965870 M * matti Hi Bertl :) 1121965889 M * matti Bertl: So it'a a morning now in Bertl Universal Time? ;) 1121965897 M * Bertl hey matti :] how does life treat you? 1121965928 M * Bertl yeah, well, I'm a little ill atm (looks like some kind of summer flu) 1121965931 M * matti s/it\'a/it\'s/ 1121965939 M * matti Bertl: :(((((((((((((((((( 1121965948 M * matti Bertl: Don't die :< 1121965962 M * Bertl lol, I'll try, tx :) 1121965965 M * matti :) 1121966081 M * matti Bertl: Want some Tetley Earl Grey tea? 1121966094 M * Bertl excellent idea! 1121966117 M * matti :) 1121966123 M * matti Bertl: Sugar? 1121966286 A * matti gives Bertl a cup of hot tea :) 1121966308 M * Bertl tx 1121966320 M * matti :) 1121969881 J * benjamin_ ~benjamin@sherpadown.net 1121970129 M * Bertl welcome benjamin_! 1121970181 M * benjamin_ hi Bertl 1121972160 N * richardw|na richardw 1121972170 M * Bertl wb richardw! 1121972196 M * richardw hi Bertl :) 1121972200 M * Bertl richardw: any news regarding the uml/skas crash? 1121972241 M * richardw nothing :( no logs (host and guest), nothing 1121972310 M * richardw maybe it is the skas + ck patch... 1121972348 M * Bertl is it reproducible? 1121972416 M * richardw no 1121972442 M * Bertl hmm hmm ... so it works sometimes? 1121972472 M * richardw yes. the uml guest was under heavy load and the host too. 1121972949 Q * eXplasm Remote host closed the connection 1121974054 J * yarihm ~yarihm@80-218-5-17.dclient.hispeed.ch 1121976295 Q * kevinp Read error: Connection reset by peer 1121976369 J * kevinp ~kevinp@ny.webpipe.net 1121979338 J * bipsen ~secret@pat.progressive.dk 1121979388 M * bipsen hi, I've tried to compile the new rc8.1 kernel - but I keep getting kernel panic's ... the old 1.9.5 kernel was working okay on the host... 1121979425 M * bipsen Should I try a vanilla kernel (without vserver patch) first, just to make sure, that it isn't the vserver patch, that is causing the problem ? 1121979485 M * bipsen Can't remember the exact kernel panic message... something about it want's to kill the init (or something like that)... How does one debug/capture the output in this situation ? 1121979654 M * FaUl bipsen: serial console 1121979677 M * FaUl bipsen: and trying a non-vserver-kernel would help 1121979798 M * bipsen FaUl: Okay, I'm no expert in debugging or similar... just attach a null-modem cable on ttyS0 ?? And what to connect at the other end ? A plain old terminal program, or another linux box running a special application ? 1121979974 M * bipsen Only problem right now is that the box i question is located at work, and I cannot access it before tomorrow morning (08:00 CET) .. So I might as well use some time now to try to figure out what to do ;-) 1121980014 M * bipsen FaUl: okay - found some info (google is a good friend ;-)): http://www.faqs.org/docs/Linux-HOWTO/Remote-Serial-Console-HOWTO.html 1121980050 M * Bertl bipsen: a vanilla kernel to test is a good idea 1121980067 M * Bertl same config options as for the vserver kernel 1121980084 M * Bertl bipsen: btw, do you get the panic on boot, or while working 1121980311 M * bipsen On boot.,.... 1121980378 M * bipsen as far as I remember, the message is something like "tried to kill init" ... I'll try a vanilla kernel tomorrow - just to see wheter that one will run or not... 1121980406 M * Bertl bipsen: that looks more like a setup issue than a real kernel issue ... 1121980824 J * eXplasm explasm@p549F7573.dip.t-dialin.net 1121980872 M * bipsen Well, I took the .config from my 1.9.5 vserver, did "make oldconfig", answered no to a lot of the new stuff (don't need a lot of joystick ddrivers etc), make, make modules, make modules install, copied kernel and system.map - did mkinitrd .. have I forgotten something 1121980874 M * bipsen ? 1121980913 M * Bertl no, sounds good, but maybe something isn't as before ... 1121980931 M * Bertl the actual panic will give a good hint I guess 1121980986 M * bipsen I'll try to see if I can get grub etc configured tomorrow for serial console ... (don't knpw whether this is default in all kernels as well - guess I'll have to check all settings) 1121981078 M * bipsen It's in times like this I wish my boss would sponsor a HP/Compaq server with iLo - so I could take complete control over it, even is it crashed ;-) 1121981130 M * Bertl well, you can do the same with much less efford/money 1121981149 M * Bertl but of course, it's nice with ILO 1121981213 M * bipsen Well - remote power off/reboot can be done using ethernet-controlled power-outlets... But I still need to be able to change the booting kernel during startup. That is where ILO would be nice 1121981274 J * Aiken ~james@tooax6-124.dialup.optusnet.com.au 1121981295 M * Bertl bipsen: I do that via serial console on most machines 1121981300 M * Bertl welcome Aiken! 1121981302 M * bipsen actually, the server in question is a Intel motherboard - a good, old SPKA4 motherboard with 2 P-III Xeon CPU's 1121981307 M * Aiken hello 1121981379 M * bipsen Bertl: okay... that just requires some kind of host or box with a bunch of serial ports... Is it possible to run output on both the monitor and serial console ? - I never use X11 anyway... 1121981406 M * Bertl yes, that works too, but it is a little more complicated 1121981428 M * Bertl but usually you are fine if the 'kernel' specific part just is via serial console 1121981443 M * Bertl you can keep the virtual logon consoles on the monitor ... 1121981494 M * bipsen okay - that will just result in the grub-loader being presented on the serial console only ? 1121981510 M * bipsen and no way to choose on the "local" monitor ? 1121981530 M * Bertl I have a patch where you can get the grub on monitor _or_ console 1121981549 M * Bertl depending on which one you use for the first input, it will then get locked to 1121981746 M * bipsen sounds really neat... Is it somewhere on the net, or could you mail it to me ? 1121981778 M * bipsen could be very handy in situations like this ;-) 1121981786 M * Bertl should be somewhere on the net ... let me check ... 1121981892 J * revenger_ ~revenger@p54B00446.dip0.t-ipconnect.de 1121981898 M * revenger_ Bertl: anwesend? 1121981936 M * Bertl yep, I'm here ... 1121981944 M * revenger_ hmm has vserver it's own tcp stack or is it just a host environment which shares the main resources? 1121981970 M * Bertl atm, (without ngnet) it shares the hosts tcp stack 1121982002 M * Bertl i.e. guests are limited to a subset of ips, but except for that no changes ... 1121982004 M * revenger_ what's ngnet? 1121982014 M * Bertl the so called next generation networking ... 1121982030 M * Bertl (which will allow for virtualized networking) 1121982032 M * revenger_ does it use the tun/tap extensions? 1121982048 M * Bertl no, because that would add too much overhead 1121982049 M * revenger_ think that could solve my problem :-) 1121982056 M * revenger_ why so? 1121982064 M * Bertl because tun/tap is userspace 1121982075 M * revenger_ ok 1121982096 M * Bertl we do not really want to send packets from a guest through the kernel into userspace, jsut to send them back into the kernel (as UML does it for example) 1121982183 M * revenger_ ok sounds like a nice design roadmap.. 1121982217 M * Bertl bipsen: just uploaded them to: http://vserver.13thfloor.at/Experimental/grub-patches/ 1121982247 M * Bertl bipsen: you probably have to adjust them to newer grubs ... 1121982311 M * bipsen Bertl: Okay, that is probably a minor issue.... I'll take a look at it, and see if it shouldn't be possible to get something up and running 1121982340 M * bipsen all 6 patches are needed ? 1121982369 Q * Aiken Quit: Leaving 1121982381 M * Bertl bipsen: no, they have different purposes ... 1121982459 M * bipsen okay - I'll tro to have a look at them..... 1121982473 Q * bipsen Quit: reboot required :-( - windows update 1121983174 J * bipsen ~secret@pat.progressive.dk 1121983214 M * bipsen Bertl: I guess the patch I need for the serial-port stuff (choosing kernel in Grub) is the grub-0.93-console.patch ? 1121983248 M * Bertl yep 1121983308 Q * yarihm Quit: Leaving 1121983539 J * Aiken ~james@tooax6-124.dialup.optusnet.com.au 1121983838 Q * Aiken Quit: Leaving 1121983930 M * revenger_ Bertl: have you ever tested port knocking? 1121984052 M * Bertl revenger_: well, not tested, but I know what it is 1121984178 M * revenger_ think I'll set up my workstation at work with that feature since I have an external IP as well 1121984197 P * bipsen 1121984675 J * Aiken ~james@tooax6-124.dialup.optusnet.com.au 1121985017 Q * benjamin_ Quit: Pwet 1121985018 Q * richardw Read error: Connection reset by peer 1121985501 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il 1121986749 J * richardw ~richard@81.189.230.20 1121986772 M * Bertl wb flock! richardw! 1121986812 M * richardw hi Bertl, my stupid internetconnection crashed :( 1121986841 M * Bertl happens now and then ... 1121987466 M * FaUl narf, why does no spam arriving if i'm waiting for it for testing my spamfilter-setup? ;-) 1121987495 M * FaUl ah, finally 1121987501 M * FaUl so, time to go to bed now 1121987517 M * Bertl hehe, good night then! 1121987518 M * richardw do you need some spam? ;-) 1121987529 M * richardw gn8 Bertl! 1121987547 M * Bertl hmm, it's FaUl going to bed I guess :) 1121987623 M * richardw foo ;) i better should go to bed ;-)