1120609559 Q * Doener Quit: Leaving 1120610517 J * Doener ~doener@p54876E3A.dip.t-dialin.net 1120610709 M * Doener hmm, seems that i messed that up :) 1120610715 M * Doener 1: 192.168.100.101/0.255.255.255 1120610874 M * Vudumen gn allz 1120610934 M * Doener night Vudumen 1120611064 M * Doener yeeha! we have on-the-fly ip address adding :) 1120611159 M * Doener what should probably also be changed is the binding code, that currently converts 0.0.0.0 to the actual address if the network context has only one ip address 1120611459 N * Bertl_oO Bertl 1120611463 M * Doener wb Bertl 1120611474 M * Bertl tx, evening everyone! 1120611488 M * Doener /tmp/ccQTSv77.s: Assembler messages: 1120611488 M * Doener /tmp/ccQTSv77.s:404: Error: suffix or operands invalid for `lea' 1120611501 M * Doener what does this tell me? 1120611514 M * Bertl that your toolchain is broken? 1120611521 M * Doener hm... 1120611548 M * Doener int ret; 1120611548 M * Doener ret = vserver(VCMD_net_add, nid, &res); 1120611548 M * Doener printf("Retval: %d\n", ret); 1120611550 M * Bertl well, the compiler produced some assembler code, your assembler did not agree with :) 1120611569 M * Doener if the printf() is moved before the vserver(), it compiles... 1120611583 M * Bertl gcc/gas version? 1120611619 M * Doener gcc version 3.3.5-20050130 (Gentoo Linux 3.3.5.20050130-r1, ssp-3.3.5.20050130-1, pie-8.7.7.1) 1120611634 M * Doener GNU assembler version 2.15.92.0.2 (i686-pc-linux-gnu) using BFD version 2.15.92.0.2 20040927 1120611660 M * Bertl okay, please upload an example code ... 1120612041 M * Bertl btw, gcc isn't that good with inline assembler ... 1120612056 M * Bertl (especially on x86 it gets some things wrong, now and then) 1120612097 M * Bertl so it might be better to put the vserver calls into separate functions, which take the correct arguments, fill them in the data structures, and execute the syscall 1120612150 M * Doener the printf is just there because it segfaulted right after the syscall 1120612361 M * Doener hm, might be library related again... can't get a testcase with a simple program to fail 1120612385 M * Doener Bertl: http://www.13thfloor.at/~doener/vserver/patches/delta-netnew-feat02.diff 1120612417 M * Doener your patch would run over the boundaries of the array if count > 4 1120612448 M * Bertl ah, right, didn't think of those evil folks out there, thanks! 1120612473 M * daniel_hozac i just looked at it again. should it really be &ip and &mask in the copy_from_user? 1120612474 M * Doener and for testing I also removed the #ifndef CONFIG_VSERVER_LEGACYNET... is that needed anyways? 1120612492 M * daniel_hozac (in vc_net_add()) 1120612506 M * Doener of course not... 1120612509 M * Doener thanks! 1120612510 M * Bertl no, was just for 'testing' 1120612528 M * Doener had some other code there earlier... will fix that... maybe that also causes the segfaults 1120612545 M * Bertl copy_from_user(&ip, vc_data.ip, vc_data.count * sizeof(*ip)) || 1120612546 M * Bertl +copy_from_user(&mask, vc_data.mask, vc_data.count * sizeof(*mask))) 1120612553 M * Bertl what's that supposed to do? 1120612573 M * daniel_hozac copy the arrays from userspace to kernelspace? 1120612582 M * Doener i replaced uint32_t ip[4] with uint32_t *ip, same for mask 1120612592 M * Doener thus i has to copy the actual addresses 1120612600 M * Bertl ahem, that doesn't make sense 1120612609 M * Bertl we do not want pointers here 1120612653 M * Doener just ip[NB_IPV4ROOT]? 1120612657 M * daniel_hozac hmm, why not? 1120612671 M * Bertl because we have to fix them up in the compatibility case 1120612680 M * Bertl and it doesn't buy us anything but pain 1120612686 M * daniel_hozac ah, right. 1120612733 M * Bertl if you don't like the 'count' (with a max of 4 for ipv4 and 1 for ipv6, which is the reason for having it size 4) 1120612756 M * Doener ah, combined thingy... 1120612758 M * Bertl then we can make it an union which carries only 1 ip at a time 1120612772 M * Doener yep, was about to suggest the same :) 1120612794 M * Bertl but I thought that 4 ipv4 ips would catch 90% of the setups in one call 1120612796 M * daniel_hozac the union is probably the least confusing of the two. 1120612866 M * Bertl well, I think it will be more confusing but as I said, I can live with that 1120612884 M * daniel_hozac why would the union be more confusing? 1120612922 M * Bertl because having uint32_t and uint32_t [4] in the same union at least confuses me :) 1120612946 M * daniel_hozac ah.. true. 1120613174 M * Doener I'm off to bed now... probably not back before thursday 1120613178 M * Doener have a good one! 1120613182 M * Bertl so the remaining question is: do we prefer a 'single' IP passing version of add/remove over an (up to 4)/1 ip version 1120613185 M * daniel_hozac good night. 1120613194 M * Bertl Doener: good night then! 1120613210 M * daniel_hozac i guess the up to 4 one is preferred. 1120613214 A * Doener votes for single addresses... 1120613250 M * Doener the count thing is IMHO quite confusing if it is limited depending on the address type... 1120613257 M * Bertl well, actually we can use the up to 4 for single addresses too, no? 1120613277 M * Bertl (i.e. just don't set count > 1 :) 1120613319 M * Doener "single address passing"... :p 1120613324 M * Doener gone now ;) 1120613426 M * Bertl otoh, we could once again forget about ipv6 (because it's probably too intrusive for non ngnet anyway) and just pass the array of all (16 currently) ipv4 addresses at once 1120613441 M * Bertl which would remove the need for a set/remove completely 1120613539 M * Bertl Hollow: still around? 1120613616 M * daniel_hozac he went to bed at 00:01 ;) 1120613647 M * Bertl ah, okay, missed that ... 1120613685 M * Bertl well, so it seems you are up-to-date, right? 1120613699 M * Bertl can you give me a short overview about the library progress? 1120613765 M * daniel_hozac i think most of the syscalls are implemented 1120613771 M * daniel_hozac s/most/all/ 1120613834 M * daniel_hozac i haven't been able to follow the library progress too closely myself. 1120613905 M * AprilDL bertl: finally got my 2.6.11 kernel to boot and have success on every line of testme.sh 1120614147 M * Bertl AprilDL: congrats! 1120616111 A * Bertl didn't want to abruptly stop any discussions :/ 1120617207 J * itamarjp lualele@200-225-242-108.dynamic.idial.com.br 1120617875 M * Bertl welcome itamarjp! 1120618922 J * eXplasm explasm@p549F7D53.dip.t-dialin.net 1120619285 M * Bertl welcome eXplasm! 1120619392 Q * baggins Ping timeout: 480 seconds 1120619744 M * itamarjp Hi Bert1 1120620028 M * Bertl ah, already thought I lost connection :) 1120623652 Q * Aiken Quit: Leaving 1120623985 J * Aiken ~james@tooax8-075.dialup.optusnet.com.au 1120624200 M * Aiken Bertl it looks like my syslogd + vserver problem could still be alpha related 1120624215 M * Bertl hmm? 1120624229 M * Aiken just changed this machine to 2.6.12.2-rc5 and it works fine 1120624234 M * Aiken athlon xp 1120624288 M * Aiken this thing has glibc 2.3.2 and the alpha has glibc 2.3.5, only difference in toolchain. everything same versions 1120624305 M * Bertl interesting ... 1120624336 M * Bertl btw, the cleaned up version of our fix (from yesterday) is here: http://vserver.13thfloor.at/Experimental/FOR-2.0/delta-alpha-fix01.diff 1120624359 M * Bertl (this should work ontop of rc5) 1120624410 M * Aiken last night I started with clean tree, applied rc5 and the deltas then made my own patch from what was done yesterday 1120624423 M * Aiken and already used it 1120624446 M * Bertl yeah, it should not change anything, it's supposed to be a cleaned up version 1120624459 M * Aiken I have saved it 1120624526 M * Aiken only complaint with the x86 version so far is I need to download new nvidia drivers 1120624543 M * Aiken driver that compiled against 2.6.11 won't against 2.6.12 :( 1120624564 M * AprilDL still using the 1.95 patch on 2.6.11 kernel - is this the best util-vserver for me (debian) util-vserver_0.30.204-5sarge2_i386.deb 1120624591 M * Bertl if you say so ... :) 1120624656 M * Bertl AprilDL: but I guess Aiken will have no luck getting that one to run on his Alpha *G* 1120624661 M * AprilDL does that mean the debian maintainer is in the channel and you don't want to say anything bad about him? 1120624693 M * AprilDL i had a red alfa convertible once. very powerful for a 4 cylinder. 1120624706 M * Bertl no, it just means that as you added 'for me', I can hardly object, no? 1120624751 M * Bertl for all others, I suggest to get a recent version (kernel patch and tools) 1120624771 M * AprilDL i could install from the 13th floor if its better actually installing from source isn't always so different from apt get in simple situations 1120624797 M * AprilDL i can't go thru the kernel patch again tonight i don't think. 1.95 is so bad? 1120624804 M * AprilDL i thought it wise to wait for 2.1 1120624813 M * AprilDL rather than 2.0rc5 1120624823 M * Bertl hmm, you like the devel releases more than the stable ones? 1120624840 M * AprilDL ya talked me into skipping 1.2 1120624843 M * AprilDL on 2.4 1120624864 M * AprilDL its hard for me to judge other than - open hosting appears to be using 1.95 1120624881 M * Bertl well, 1.9.x (devel release), 2.0.x (stable release), 2.1.x (devel release) 1120624888 M * AprilDL :) 1120624917 M * Bertl very similar to ... hmmm, let me think, what was it? ... ah, yeah, the kernel :) 1120624923 M * AprilDL ok if its important that i go to 2.0rc5 instead of wait for 2.0.... 1120624928 M * AprilDL i mean i can't do this every couple of days 1120624946 M * AprilDL as if i would know about kernels. 1120624949 M * Bertl no, it's not critical, and IMHO if it works for you right now, why change it 1120624953 A * AprilDL checks to see if booted from channel 1120624964 M * AprilDL ok good 1120624974 M * AprilDL so you have no horrible warnings about util-vserver_0.30.204-5sarge2_i386.deb 1120624987 M * Bertl what you should do is to question the quality of debian/sarge packages 1120625002 M * AprilDL this is not my usual job but i feel i'm gaining useful experience 1120625009 M * Bertl because at least regarding linux-vserver they have a terrible quality 1120625040 M * Bertl and you might check now and then, what changed/was fixed/was added in the mainline releases 1120625049 M * AprilDL is sarge = unstable? its late here. my bedtime. 1120625069 M * Bertl sarge is now called 'stable' but I'd call it 'broken' :) 1120625090 M * AprilDL oh then i hope you like unstable better 1120625113 M * Bertl but more important, this is actually not the place to discuss sarge issues, unless they are linux-vserver related 1120625123 M * AprilDL the debian channel says unstable is now once again really unstable. but i dont' want to take up space here saying meaningless stuff. i should go to sleep 1120625131 M * AprilDL right 1120625151 M * Bertl and as the current maintainer (of linux-vserver) is _not_ listening here ... there is no point in bringing up issues either ... 1120625164 M * AprilDL yes but its not a good sign 1120625168 M * AprilDL i dont' hear good things 1120625257 M * Bertl so from my point of view, I basically do not want to be bothered with 'debian' issues, I want to deal with linux-vserver issues, and if the debian maintainer is modifying/cripling/whatever the mainline stuff, that is not really my business ... 1120625338 M * Bertl to make it short: if debian 'whatever' works for you (with package XY), then that's pretty fine and we are happy with you, but if it doesn't work, get the mainline version and show us that it is a linux-vserver issue, not a debian issue ... 1120625476 M * Bertl that said, I'm off to bed to avoid further rant against distros ... 1120625487 M * Bertl have a nice whatever everyone ... cya tomorrow! 1120625494 N * Bertl Bertl_zZ 1120626507 Q * jonsmel_zZ uranium.oftc.net keid.oftc.net 1120626692 M * _are_ AprilDL: I use the vserver tools provided by debian sarge, but I would always use the most recent patches/kernel version from 13th-floor. But then again I have 1 dual Opteron vserver host and x86_64 is quite broken in 2.6 < ... 1120626692 M * _are_ 2.6.11.11, aka race conditins leading to a system freeze with heavy IO load. 1120626698 J * jonsmel_zZ ~jscottorn@209.33.206.3 1120627917 Q * itamarjp Quit: 1120628273 Q * Aiken Quit: Leaving 1120632722 J * sukria ~sukria@213.223.184.194 1120634194 J * erwan_ho ~erwan@konilope.dyndns.org 1120634264 J * Doener` ~doener@p54874EAE.dip.t-dialin.net 1120634697 Q * Doener Ping timeout: 480 seconds 1120634995 Q * erwan_ho Remote host closed the connection 1120635397 Q * _are_ Quit: bbl 1120635606 Q * sukria Quit: coming back soon 1120635747 J * sukria ~sukria@www.sukria.net 1120635887 J * prae ~prae@ezoffice.mandriva.com 1120636592 Q * [maple] Read error: Connection reset by peer 1120636652 J * [maple] ~maple@strace-nat0.ua-online.net 1120637421 Q * [maple] Read error: Connection reset by peer 1120637465 J * [maple] ~maple@strace-nat0.ua-online.net 1120640408 J * baggins baggins@kenny.mimuw.edu.pl 1120641262 M * Hollow morning 1120641479 M * daniel_hozac morning 1120642027 M * DaPhreak lo Hollow, daniel_hozac 1120642090 M * Hollow lo DaPhreak 1120642194 M * DaPhreak Hollow: i'm gonna update *my* vserver-patches to 2.6.12 ... ;P hate it to wait for the grsecurity-patch 1120642213 M * Hollow heh k.. use rc5 plz 1120642224 M * DaPhreak yeah ... 1120642240 M * DaPhreak well it's actually finished .. just need to test it later (at home) 1120642299 M * Doener` morning guys 1120642307 M * DaPhreak morning Doener` 1120642326 M * Doener` hm, so we got 2.6.13-rc2.5? ;) 1120642331 M * DaPhreak eh ?! 1120642344 M * DaPhreak *g* nah .. only 6.13-rc3 ;) 1120642379 M * Doener` 6.13? that's a developers kernel then... not interested ;) 1120642400 M * Hollow 2.6.12.2-vs2.0-rc5 1120642402 M * DaPhreak heh *g* well you actually running 6.12-vs2.0rc5 ;) 1120642441 A * Doener` must be living in the past... 1120642453 M * DaPhreak yeah .. me too 1120642460 M * DaPhreak still running 2.6.11.12 ;P 1120642500 M * Doener` i got some bastard between 2.6.12.2-vs2.0-rc5 and ..-vs2.0-rc5.4 1120642540 M * DaPhreak hmm ... :) well then my new must be something like vs2.0-rc5.1 ;P 1120642805 J * _are_ ~are@gateway-dsl.lihas.de 1120642861 M * _are_ hi 1120642867 M * Doener` morning _ag_ 1120642880 M * Doener` erhm s/_ag_/_are_/ 1120642895 M * _are_ ah, almost ;) 1120644958 Q * cryo Read error: Connection reset by peer 1120645066 J * cryo ~say@212.86.243.154 1120649159 Q * nox Remote host closed the connection 1120649205 J * nox ~nox@noxlux.de 1120649756 Q * nox Remote host closed the connection 1120649792 J * nox ~nox@noxlux.de 1120650960 M * Hollow Doener`: svn up plz 1120650969 M * Hollow big cleanup 1120650972 M * Hollow in tools 1120651003 M * daniel_hozac is the libvserver team really a legal entity? ;) 1120651031 M * Hollow depends on your definition of 'legal' ;) 1120651314 Q * nox Quit: Changing server 1120652105 J * pflanze ~chris@62.12.148.89 1120652220 M * pflanze Hello 1120652245 M * pflanze I'm wondering where the 2.0preX release is linked 1120652277 M * daniel_hozac http://vserver.13thfloor.at/Experimental/ 1120652289 M * daniel_hozac http://vserver.13thfloor.at/Experimental/patch-2.6.12.2-vs2.0-rc5.diff is the most recent one. 1120652311 M * pflanze yup thanks. 1120652512 M * pflanze I see. It's not linked under the Downloads section. Nothing references "experimental" there explicitely. If I knew whether Experimental should be pointed out I would edit the wiki page. 1120652589 M * meebey hm I am using now the new util-vserver, strange layout :) 1120652596 M * meebey trying to migrate 1120652855 M * BWare there is a 'migration script' available in the mailarchives 1120652867 M * meebey which month? :) 1120652869 M * BWare did not get around posting it on the wiki ;) 1120652871 M * BWare pfff 1120652877 M * BWare good question ;) 1120652896 M * BWare june 1120652904 M * BWare according to my sent mail log 1120652931 M * BWare From: Dennis Roos 1120652931 M * BWare To: vserver@list.linux-vserver.org 1120652932 M * BWare Subject: Re: [Vserver] Migration from legacy utils to the new version 1120652932 M * BWare Date: Mon, 06 Jun 2005 15:26:26 +0200 1120653120 M * meebey thanks 1120653230 J * nox ~nox@noxlux.de 1120654604 M * meebey BWare: hm the script looks interesting 1120654610 M * meebey not sure if I trust it 1120654648 M * meebey caps is comment out btw 1120654655 M * meebey commented 1120654967 M * meebey I will try it anyhow, porting the caps is easy 1120654986 M * meebey oohh wait, I have to patch the tools now 1120654996 M * meebey else VPN will not work anymore :) 1120655010 M * meebey Doener`: ping 1120655070 M * meebey Doener`: the proc write patch I gave me once, I ported it for each vserver and util-vserver version, but now we have this new util-vserver programs, could you help me porting the patch? 1120655219 M * BWare meebey: Enjoy the script.. it saved me loads of time when I had to convert lots of servers :) 1120655268 M * meebey :) 1120655285 M * meebey hmmm klemme think 1120655291 M * meebey 30 servers AFAIK 1120655298 M * meebey about 5 vservers per server 1120655370 M * DaPhreak meebey: could you give me a pointer to your old patch ?! :) 1120655439 M * meebey lemme upload it, one sec 1120655454 M * meebey I am running freeswan/openswan inside a vserver with it 1120655478 M * DaPhreak your vpn stuff (you mentioned on the ml) 1120655493 M * meebey it's not very secure because of proc write access, but I need it mostly the for modularness 1120655695 M * meebey DaPhreak: http://www.meebey.net/temp/vserver/ 1120655768 M * DaPhreak *uh* for 2.4 *g* 1120655793 M * meebey yes 1120655832 M * meebey not ready yet for 2.6 1120655841 M * meebey isdn, vpn, etc must be tested first 1120655852 M * meebey and vserver of course 1120656020 N * jonsmel_zZ jonsmel 1120656030 M * DaPhreak lo jonsmel 1120656039 M * jonsmel hi there 1120656043 M * meebey generic.sh? 1120656050 M * meebey it looks bogus to me 1120656059 M * meebey galilei:/etc/vservers/dev/scripts# cat generic.sh 1120656059 M * meebey #!/bin/sh 1120656059 M * meebey case "$1" in 1120656067 M * meebey thats the whole script 1120656077 M * meebey is that on purpose? 1120656101 M * DaPhreak meebey: well that patch (at least the kernel side) shouldn't be a big deal .. since the patch does only 2 things .. 1120656136 M * meebey DaPhreak: only if you know vserver internals :) 1120656148 M * DaPhreak heh, well or if you find something similar ;) 1120656153 M * meebey I know what vserver does but not how 1120656167 M * DaPhreak Doener`: ? 1120656240 Q * Vudumen Ping timeout: 480 seconds 1120656246 M * DaPhreak those #define VXF_INFO_ in include/linux/vserver/context.h, what are these 0x00000001's ? are the binary/hex values ? 1120656311 M * DaPhreak s/the/they/ 1120656402 M * meebey int 1120656410 M * meebey #define VX_INFO_LOCK 1 /* Can't request a new vx_id */ 1120656416 M * meebey #define VX_INFO_PROC_ACCESS 128 /* Allow the context to mess with the procfs */ 1120656420 M * meebey etc 1120656425 M * DaPhreak in 2.4 ... 1120656430 M * meebey yes 1120656443 J * Vudumen vudumen@perverz.hu 1120656452 M * DaPhreak if you look at the 2.6 ones .. those are something different ;) at least they look different that int 1120656468 M * meebey 0x01 is hex 1120656471 M * DaPhreak #define VXF_INFO_LOCK 0x00000001 1120656525 M * meebey hm I am missing something in the new structure 1120656536 M * meebey how can I set that the vserver should not boott? 1120656551 M * meebey ONBOOT=false that was in the old version 1120656566 M * meebey ONBOOT=no even 1120656714 M * DaPhreak hmm .. by default the do not boot, until you add "default" to /etc/vservers//apps/init 1120657265 M * meebey ah ic 1120657271 N * Bertl_zZ Bertl 1120657276 M * DaPhreak morning Bertl ;) 1120657277 M * meebey then is the migration script not complete :) 1120657286 M * Bertl morning folks! 1120657286 M * meebey it addded apps/init 1120657289 M * meebey hixa BWare 1120657291 M * meebey er 1120657294 M * meebey hiya Bertl 1120657295 Q * [maple] Read error: Connection reset by peer 1120657301 J * [maple] ~maple@strace-nat0.ua-online.net 1120657317 M * meebey and the network migration looks buggy to me 1120657323 M * meebey galilei:/etc/vservers/dev# cat interfaces/0/dev 1120657323 M * meebey 169.254.0.253 1120657323 M * meebey galilei:/etc/vservers/dev# cat interfaces/0/ip 1120657323 M * meebey 169.254.0.253 1120657323 M * meebey galilei:/etc/vservers/dev# cat interfaces/0/mask 1120657326 M * meebey 169.254.0.253 1120657329 M * meebey :) 1120657339 M * DaPhreak heh, a bit ;) 1120657345 M * meebey dev is eth0? 1120657352 M * Bertl ah, _another_ migration script? 1120657357 M * DaPhreak well yeah (if it exists on the host) 1120657360 M * meebey IPROOTDEV=eth0 1120657363 M * DaPhreak Bertl: no, BWare's 1120657375 M * meebey thats what I had 1120657396 M * DaPhreak yeah .. then its eth0 1120657477 M * meebey hmm I think I should fix the script 1120657480 M * Bertl well, IIRC, I suggested to source the original config ... 1120657484 M * meebey else I a lot of work 1120657494 M * meebey +have 1120657518 M * meebey Bertl: good idea its only variables 1120657523 M * Bertl it should then be easy to formulate some fallback conditions 1120657566 M * Bertl (i.e. split the interface/address/mask, if interface or mask is missing output the defaults 1120657596 M * Bertl but be careful, if there is no interface/mask set at all, you have to add nodev 1120657801 M * meebey hm 1120657839 M * BWare as I stated in the original post: Use at your own risk :) 1120657855 M * meebey BWare: thats why I am tesitng it 1120657873 M * meebey BWare: the page is also helpful 1120657878 M * meebey ugly colors but helpful 1120657890 M * BWare meebey: Try a different style sheet ;) 1120657919 M * BWare I always use 'no stylesheet' on the flower page 1120657931 M * meebey ah percet 1120657933 M * meebey no 1120657940 M * meebey no styleis best 1120657942 M * meebey :) 1120658010 M * meebey erm 1120658015 M * meebey where is the priority? 1120658066 M * meebey # Obtain all settings from the Linux-VServer using old configfile 1120658066 M * meebey source /etc/vservers/${VSNAME}.conf 1120658075 M * meebey Bertl: it does it 1120658135 M * Bertl it does not what? 1120658160 M * meebey 15:44:40 < Bertl> well, IIRC, I suggested to source the original config ... 1120658171 M * meebey Bertl: I read that like it does not do that :) 1120658219 M * Bertl ah, okay, sorry ... probably it's doing fine, didn't look at it closer 1120658241 M * BWare meebey: It is a script written to migrate the functionality of my own environment, I did not check all configuration scenario's 1120658266 M * Bertl BWare: which is perfectly fine! don't worry ... 1120658299 M * meebey BWare: its a great start for me, no problem, I just need to add some bits 1120658324 M * meebey generic.sh is strange, its a header of a real script? 1120658324 M * BWare Go ahead, change it, use it and look at the topic ;) 1120658358 M * BWare generic.sh was needed in our environment 1120658364 M * meebey ah ic 1120658374 M * meebey I thought its some vserver magic 1120658385 M * BWare it consists of a couple of functions that are called from the various stages 1120658398 M * meebey ic 1120658411 M * BWare (calling vdlimit, setting routes/rules/etc.) 1120658417 M * meebey I need that too, a firewall restart script 1120658435 M * meebey maybe the bug is fixed in util-vserver now 1120658458 M * meebey calling vserver foo running returned 0 when called from post-stop 1120658478 M * meebey and 1 when called from post-start 1120658485 M * BWare Heh, there are a gazillion reasons for that 1120658516 M * meebey ah little (sleep 2; foo)& solved that but is ugly 1120658534 M * BWare might be that some processes were not killed, and interface could not be removed, an additional CR/LF at some config file, etc.,etc. 1120658551 M * meebey no it was really stopped 1120658560 M * meebey but just after post-stop finishes :) 1120658570 M * meebey it's post-stop so it should be stopped by that time 1120658587 M * BWare Ah, there was some discussion about something similar recently 1120658596 M * meebey its _post_-stop not on-stop 1120658601 M * BWare so it's probably fixed in the new tools 1120658606 M * meebey yeah 1120658609 M * meebey I will test it 1120658669 M * BWare Bertl: I have experienced some crashes lately that might be related to vservers 1120658697 M * BWare The problem is that the whole machine hangs so debugging is kinda problematic 1120658713 M * meebey "case $1 in") 1120658717 M * meebey bug 1120658722 M * meebey should be \$1 1120658737 M * meebey thats why the case thing was left in generic.sh 1120658853 M * BWare post your modifications to the mailing list, or to the wiki if you want to, so others can use them as well 1120658869 M * BWare brb... caffeine input required 1120659379 M * meebey so where is priority? 1120659459 M * meebey hehe there is postpost-stop now 1120659461 M * meebey :) 1120659509 M * meebey hhmm depends seems to solve that one 1120659510 M * meebey nice 1120659616 Q * [maple] Remote host closed the connection 1120660025 M * Bertl okay, have to leave now ... back later ... 1120660030 N * Bertl Bertl_oO 1120660109 J * pmenier ~votremail@ACaen-202-1-2-5.w193-251.abo.wanadoo.fr 1120660109 J * [maple] ~maple@strace-nat0.ua-online.net 1120660130 M * pmenier hello 1120660155 M * pmenier is there anybody who could help me to make vserver run 1120660185 M * jonsmel just ask and ye shall receive 1120660190 M * pmenier ok 1120660230 M * pmenier i(ve installed a kernel 2.6.12.2 with kernel patch-2.6.12.2-vs2.0-rc5.diff 1120660260 M * pmenier then i downloaded util-vserver from linux-vserver. i compile it. All seems ok 1120660283 M * pmenier but when i launch vserver vweb1 start i take Can't set ipv4 root (invalid argument) 1120660307 M * pmenier i run on debian with a kernel from kernel.org 1120660340 M * pmenier on the same machine with a kernel 2.6.11.11 no problem 1120660360 M * pmenier excuse but my english isn't perfect 1120660388 M * Hollow pmenier: did you enable legacy networking? 1120660401 M * pmenier how do i do that ? 1120660427 M * Hollow CONFIG_VSERVER_LEGACYNET=y 1120660429 M * Hollow in kernel config set 1120660431 M * Hollow argh 1120660442 M * Hollow other way round ;) 1120660480 M * pmenier grep CONFIG_VSERVER_LEGACYNET .config => CONFIG_VSERVER_LEGACYNET=y 1120660486 M * Hollow hm... 1120660624 M * pmenier when i launch vserver vweb1 start the iface is up !!! eth0:vweb1 ok 1120660877 Q * alexx Read error: Connection reset by peer 1120660919 M * pmenier ... 1120660948 J * alexx ~alexx@82.225.136.176 1120661045 M * pmenier i leave 5mn to take a coffee 1120661319 M * pmenier hello i'm back 1120661490 M * pmenier i've wanted to try the debian package but it's in conflict with vpopmail 1120661528 M * pmenier another strange thing : i've tried the same config in a vmware : all works fine 1120662010 M * _are_ pmenier: afaik the consens in here is the debian sarge package is broken 1120662045 M * _are_ works fine for me, but I have no vpopmail installed. :-> 1120662065 M * pmenier yes it's why i wanted to compile it from linux-vserver.org 1120662098 M * _are_ yes, compiled fine here whenever i needed it, I always take the latest version avaliable 1120662138 M * pmenier it's what i wanted to do so 1120662174 M * pmenier but i don't really understand why it works fine in a vmware and not on a real machine 1120662204 M * _are_ what architecture do you run on? 1120662209 M * eyck yeah, real machines sux. 1120662228 M * pmenier celeron 2.4 1Go Ram 1120662261 Q * monrad Quit: Leaving 1120662293 M * pmenier it was working nice from 2.4.25 to 2.6.11.11 1120662350 M * kevinp anyone get iptables to work inside a vserver guest? 1120662446 M * id kevinp, you have to control iptables from the vserver host 1120662499 M * kevinp ok, thanks 1120662753 M * micah here is why the package conflicts with vpopmail-bin: 1120662758 M * micah * we have to conflict with vpopmail-bin because we have a 1120662758 M * micah file with the same name in the package 1120662771 M * micah http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=254277 1120662771 M * pmenier yes i know it's the vkill prgram 1120662847 M * SiD3WiNDR thats a lame reason 1120662864 M * SiD3WiNDR :P 1120662974 M * pmenier yes but about the error Can't set ipv4 root (invalid argument) ..... 1120662983 M * pmenier no ideas ??? 1120663166 M * _are_ so, no idea here, missing /proc perhaps? 1120663179 M * pmenier no /proc is ok too 1120663190 M * _are_ in the vserver? 1120663226 M * id /etc/init.d/vprocunhide on the host 1120663396 M * pmenier no i don't need because i didn't activate vserver/proc/Security in kernel 1120663577 M * meebey # style 1120663577 M * meebey Contains the init-style. 1120663582 M * meebey what style do exist? 1120663591 M * micah SiD3WiNDR: yeah, I agree, there must be a better solution 1120663594 M * meebey the migration script used "default" which does not exist 1120663614 M * micah pmenier: you probably want to activate that 1120663634 M * pmenier i work without this option on the previous kernel 1120663647 M * pmenier CONFIG_VSERVER_PROC_SECURE is not set 1120663657 M * SiD3WiNDR micah: then again, you should have anything on the host system except very basic tools and ssh and stuff :)\ 1120663688 M * micah SiD3WiNDR: I would agree to that also :) 1120663705 M * micah pmenier: why do you not want proc security? 1120663762 M * pmenier i just wanted to try vserver without too much work for it runs 1120663908 M * _are_ meebey: encountered the init style problem last week, her eis the result: http://linux-vserver.org/InitStyles 1120663942 M * meebey _are_: thanks 1120664036 M * pmenier micah: the result of df -hTa : none proc 0 0 0 - /opt/vservers/vweb1/proc 1120664252 M * meebey so how does the new vserver structure decide if it will start the vserver or not? 1120664471 M * meebey I like .defaults/ 1120664483 M * meebey that way I can restarts the firewall for each vserver globally 1120664629 M * meebey the network setup I lke too in the new util-vserver 1120664645 M * meebey now I can use ppp0 and such with nodev :) 1120665132 M * _are_ meebey: /etc/vservers//apps/init/mark contains a keyword, 'default' is started by /etc/init.d/vserver-default, adjust as you like 1120665164 M * _are_ and if I remember right, the new tools support vserver dependencies, too, though I haven't used them 1120665640 Q * prae Quit: Execute Order 69 ! 1120665842 J * meddochat ~meddochat@p508F6509.dip.t-dialin.net 1120665861 M * meebey _are_: so its only stats them when the mark efault is set? 1120665872 M * meddochat hi 1120665872 M * meebey starts them on host boot 1120665894 M * meddochat when i want to use vserver, do i need a physical network device for each virtual server instance? 1120665925 M * pmenier no. It creates virtual iface on the fly 1120665941 M * pmenier except for me :-( now... 1120665962 M * meddochat thanks for answering this question pmenier 1120665991 M * pmenier the iface is as : eth0: 1120666165 M * _are_ meebey: the script vserver-default only starts those vservers that have the mark 'default' 1120666171 M * _are_ at boot time, that is 1120666201 Q * meddochat Quit: 1120668257 N * jonsmel jonsmel|lunch 1120668324 J * lafeuil ~lafeuil@abo-0-229.modems.semnet.tm.fr 1120668337 M * lafeuil hi 1120668389 M * lafeuil I have a question about the file newvserver-var. It always uses in the last version of vserver ? 1120668494 J * Loki|muh loki@satanix.de 1120668757 Q * Loki|muh_ Ping timeout: 480 seconds 1120668757 Q * Loki|muh Read error: Connection reset by peer 1120668782 J * Loki|muh loki@satanix.de 1120669254 Q * pmenier Quit: >>[www.F-IRC.com - Client IRC en français]>> 1120669333 P * lafeuil 1120671553 J * ShoX0Shadow ~test@0x503e21ac.arcnxx10.adsl-dhcp.tele.dk 1120671553 Q * Pazzo Read error: Connection reset by peer 1120671629 P * ShoX0Shadow 1120672172 M * pflanze What is BME? 1120672282 M * DaPhreak Bind-Mount-Extension pflanze 1120672315 M * pflanze Does this allow mount --bind for vserver users? 1120672346 M * DaPhreak nope .. it allows -o bind,ro for example .. 1120672354 M * pflanze ah 1120673815 J * prae ~prae@sherpadown.net 1120676368 Q * _are_ Quit: bbl 1120676491 J * din ~din@24.144.40.24 1120678310 N * jonsmel|lunch jonsmel 1120679697 M * Greek0 well, -o bind,ro can be done with an unpatched kernel too, it just doesn't work as expected 1120679718 M * Greek0 (the mount is shown as ro, but you can still write at it) 1120679992 M * Greek0 hmm. I wonder where my mails to the vserver mailing list got lost.. 1120679999 M * SiD3WiNDR I thinki you can drop the "as expected" then :p 1120680124 M * Greek0 well, you can do it, and it looks like it would work in the mtab, it just doesn't prevent writing, or allow writing, or whatever you specified. 1120680139 M * Greek0 that is you can do the same to the bind mount that you could have done to the original mountpoint 1120680182 M * Greek0 anyone subscribed to the vserver ML ever got a mail from "Christian Aichinger " 1120680185 M * Greek0 ? 1120680212 M * daniel_hozac nope. 1120680225 M * daniel_hozac http://www.paul.sladen.org/vserver/archives/200507/ usually has the messages pretty instantly. 1120680241 M * Greek0 is subscription required for post rights on vserver@lists.linux-vserver.org? 1120680275 M * daniel_hozac there's a grey-listing filter at least. 1120680341 M * Greek0 yup. my mails surely looked like spam 1120680346 M * Greek0 plaintext with diff attatched.. 1120680386 M * daniel_hozac did you get a confirmation email? 1120680516 M * Greek0 I subscribed just now, got confirmation, ack'ed, resent my messages 1120680989 J * _are_ ~are@dsl-084-056-129-203.arcor-ip.net 1120681102 M * _are_ hi 1120682123 M * Greek0 hi _are_ 1120682215 M * Greek0 does anyone have an estimate how long it takes for a mail posted to vserver@lists.linux-vserver.org to be delivered? 1120682302 M * _are_ uhm, usually <10m, I think 1120682324 M * _are_ but it is some time since I last tried 1120682379 M * Greek0 *grml* I suspect that the list manager is still eating my emails 1120682811 M * SNy *yumm* 1120682829 M * SNy tasty :p 1120682837 Q * pflanze Ping timeout: 480 seconds 1120682975 M * Greek0 SNy: I really doubt if my emails are tasty.. english is not my native language, so perhaps the problem is not that the ml manager eats my mails, but that it chokes on the bad spelling ;) 1120683048 M * eyck hmm, 1120683054 M * eyck how does Bertl sleep? 1120683070 M * eyck he's on Uberman schedule? 1120683094 M * Greek0 it's 23.00 at his/my timezone, I think he should get up soon ;) 1120683802 Q * baggins Ping timeout: 480 seconds 1120683975 M * Hollow evening 1120684175 P * kevinp Leaving 1120685091 J * are|lunch ~are@dsl-084-056-155-155.arcor-ip.net 1120685237 J * Aiken ~james@tooax6-040.dialup.optusnet.com.au 1120685339 J * monrad ~monrad@213083190130.sonofon.dk 1120685490 Q * _are_ Ping timeout: 480 seconds 1120685818 J * baggins baggins@kenny.mimuw.edu.pl 1120686576 Q * virtuoso Read error: Connection reset by peer 1120686581 J * virtuoso ~s0t0na@80.253.205.251 1120686622 Q * din Quit: 5:00 1120687755 N * Bertl_oO Bertl 1120687793 M * Bertl evening folks! 1120687815 M * are|lunch Hi Bertl 1120687822 N * are|lunch _are_ 1120688318 M * Bertl hey folks! 1120688375 M * baggins hi bertl 1120688396 M * baggins Bertl: I think my problem is not with kernel but with vrpm 1120688397 M * Bertl evening Mr Baggins! 1120688422 M * Bertl hmm ... please elaborate! 1120688479 J * pflanze ~chris@pop-zh-12-2-dialup-44.freesurf.ch 1120688563 M * Bertl welcome pflanze! 1120688576 M * pflanze Hi Bertl 1120688658 M * baggins Bertl: I did some obvious tests, like chcontext chgrp, chown from inside vserver and all of it works as expected 1120688699 Q * neofutur Quit: leaving 1120688702 M * baggins the only time I have incorrect uids is when I install packages with vrpm with external package menagement 1120688718 J * neofutur ~neofutur@neofutur.net 1120688738 M * Bertl baggins: hmm ... that's a strong indication ... and a good hint, btw ... 1120688743 M * baggins I even installed debian using debootstrap and it also worked perfectly 1120688754 M * Bertl wb neofutur! 1120688774 M * Bertl baggins: you are using 2.6.12.2-vs2.0-rc5 right now? 1120688787 Q * prae Quit: Pwet 1120688833 Q * pflanze Quit: good night 1120688836 M * baggins Bertl: not now but I can confirm it with 2.6.12.2-vs2.0-rc5 when I get to work 1120688855 M * Bertl ok, np, what version did you use/try? 1120688908 M * baggins 2.6.11.12 vs2.0-rc4 with delta-xfstag-fix01.diff applied 1120688925 M * Bertl okay, you have such a system at hand right now? 1120688948 M * baggins yes, but can't reboot it 1120688968 M * Bertl np, I'd like you to do some tests for me, would that be possible? 1120688987 M * baggins ok 1120689035 M * Bertl give me a second to get my virtual test system in shape, and we do it in sync ... to compare results ... 1120689597 Q * brc Ping timeout: 480 seconds 1120689787 M * Bertl baggins: okay, I'm ready, how about you? 1120690201 J * brc bruce@201008088244.user.veloxzone.com.br 1120690206 M * Bertl wb brc! 1120690413 M * baggins Bertl: sorry for the delay, someone is playing with univ backbone 1120690420 M * Bertl np 1120690435 M * baggins i'm ready 1120690466 M * Bertl okay, I assume you have a tagxid mounted xfs? on /vservers, right? 1120690560 Q * brc Quit: No windows for this server 1120690604 M * baggins that's right 1120690619 M * Bertl okay, let's make a directory there called TEST 1120690631 M * baggins done 1120690641 M * Bertl touch /vservers/TEST/xid-0 1120690646 M * Bertl chcontext --xid 255 touch /vservers/TEST/xid-255 1120690650 M * Bertl chcontext --xid 256 touch /vservers/TEST/xid-256 1120690656 M * Bertl ls -lna /vservers/TEST 1120690672 M * Bertl (I'm only interested in the xid-* files) 1120690724 M * baggins -rw------- 1 0 0 0 Jul 7 00:58 xid-255 1120690728 M * baggins -rw------- 1 0 0 0 Jul 7 00:58 xid-256 1120690758 M * Bertl hmm, one is missing, but that's not critical yet :) 1120690759 Q * Aiken Quit: Leaving 1120690767 M * baggins -rw------- 1 0 0 0 Jul 7 00:57 xid-0 1120690771 M * baggins that one? 1120690774 M * Bertl yep 1120690783 N * jonsmel jonsmel_zZ 1120690791 M * Bertl night jonsmel_zZ! 1120690798 M * Bertl chown 20000000 /vservers/TEST/xid-0 1120690798 M * jonsmel_zZ night, night 1120690803 M * Bertl chcontext --xid 255 chown 20000000 /vservers/TEST/xid-255 1120690807 M * Bertl chcontext --xid 256 chown 20000000 /vservers/TEST/xid-256 1120690812 M * Bertl ls -lna /vservers/TEST 1120690828 M * baggins -rw------- 1 16777214 0 0 Jul 7 00:57 xid-0 1120690832 M * baggins -rw------- 1 16777214 0 0 Jul 7 00:58 xid-255 1120690836 M * baggins -rw------- 1 16777214 0 0 Jul 7 00:58 xid-256 1120690880 M * Bertl chgrp 20000000 /vservers/TEST/xid-0 1120690886 M * Bertl chcontext --xid 255 chgrp 20000000 /vservers/TEST/xid-255 1120690890 M * Bertl chcontext --xid 256 chgrp 20000000 /vservers/TEST/xid-256 1120690902 M * Bertl ls -lna /vservers/TEST 1120690907 M * baggins -rw------- 1 16777214 16777214 0 Jul 7 00:57 xid-0 1120690912 M * baggins -rw------- 1 16777214 16777214 0 Jul 7 00:58 xid-255 1120690914 M * baggins -rw------- 1 16777214 16777214 0 Jul 7 00:58 xid-256 1120690933 M * Bertl hmm .. that looks pretty sane so far ... 1120690970 M * Bertl lsxid /vservers/TEST 1120690996 M * baggins 0 /vservers/TEST 1120690999 M * baggins 0 /vservers/TEST/xid-0 1120691002 M * baggins 0 /vservers/TEST/xid-255 1120691005 M * baggins 0 /vservers/TEST/xid-256 1120691016 M * Bertl aha .. that's not correct ... 1120691035 M * Bertl 0 /vservers/TEST/xid-0 1120691035 M * Bertl 255 /vservers/TEST/xid-255 1120691035 M * Bertl 256 /vservers/TEST/xid-256 1120691050 M * Bertl you sure you mounted the /vservers with tagxid? 1120691065 M * Bertl (check with cat /proc/mounts) 1120691071 M * baggins /dev/loop1 on /vservers type xfs (rw,tagxid) 1120691105 M * Bertl is that what 'mount' says or /proc/mounts? 1120691116 M * baggins mount, /proc/mounts: 1120691119 M * baggins /dev/loop1 /vservers xfs rw,tagxid 0 0 1120691135 M * Bertl hmm ... 1120691155 M * baggins I can try all of this on ext3 if you want 1120691173 M * Bertl okay, let's do that, I'll try here with xfs (was ext2/ext3 now) 1120691201 M * Bertl ext3 on 64bit alpha, ext2 on x86 1120691735 M * Bertl hah! I can confirm that xfs doesn't show the xid correctly here too 1120691780 M * baggins on ext3 everything is correct 1120691799 M * Bertl okay, but you reported the uid/gid issues on ext3 too, right? 1120691808 M * baggins right 1120691865 M * Bertl hmm, okay, let's investigate a little further ... 1120691877 M * Bertl do you have vserver debugging compiled in? 1120691886 M * baggins I'm afraid not 1120691906 M * Bertl okay, unfortunate, but hey that's life ... 1120691931 M * baggins I didn't expect problems, I had it tested :) 1120691950 M * baggins on a different arch :/ 1120691985 M * Bertl I don't think that the arch is the issue 1120692004 M * Bertl so far the alpha (64bit) and the x86 behaves exactly the same 1120692033 M * Bertl let's try some other things ... 1120692213 M * Bertl chcontext --xid 256 --secure echo "test" >/vservers/TEST/xid-255 1120692256 M * Bertl a) does that succeed? b) do you get a kernel log entry? 1120692305 M * baggins a) yes 1120692308 M * baggins Jul 7 01:23:47 duch kernel: /sbin/vshelper: (startup 256) returned sync with 256 1120692313 M * baggins Jul 7 01:23:47 duch kernel: /sbin/vshelper: (shutdown 256) returned sync with 256 1120692346 M * baggins that's on ext3 1120692363 M * Bertl hmm, okay, does lsxid show something for TEST? 1120692378 M * baggins same for xfs 1120692395 M * Bertl you mean all have xid=0, right? 1120692411 M * baggins yes, no change 1120692457 M * Bertl chxid -c 255 /vservers/TEST/xid-255 1120692470 M * Bertl lsxid /vservers/TEST/xid-255 1120692471 M * baggins what I saw in the vserver kernel patch there is no support there for tagxid on XFS, mount options is there, but xid never gets to disk 1120692488 M * Bertl really? 1120692498 M * baggins 255 /vservers/TEST/xid-255 1120692512 M * baggins hmm, maybe I'm wrong? 1120692530 M * Bertl well, it happens through the uid/gid 1120692558 M * Bertl ls -lna /vservers/TEST/ 1120692568 M * baggins -rw------- 1 16777214 16777214 0 Jul 7 00:57 xid-0 1120692571 M * baggins -rw------- 1 16777214 16777214 33 Jul 7 01:26 xid-255 1120692574 M * baggins -rw------- 1 16777214 16777214 0 Jul 7 00:58 xid-256 1120692600 M * Bertl okay, let's umount and remount the xfs (with tagxid) 1120692622 M * baggins done 1120692635 M * Bertl lsxid /vservers/TEST/ 1120692646 M * Bertl I expect the xid tagging to be gone again ... 1120692653 M * baggins you're right 1120692654 M * baggins 0 /vservers/TEST/xid-0 1120692658 M * baggins 0 /vservers/TEST/xid-255 1120692661 M * baggins 0 /vservers/TEST/xid-256 1120692697 M * Bertl now the same is supposed to work on ext3 1120692711 M * Bertl and in my tests, both 32 and 64 bit it works there quite fine 1120692728 M * Bertl the XFS seems broken, and I'll look into it right now ... 1120692807 M * Bertl could you try some permutations of the commands we used on ext3 or maybe some 'simple' command sequence from a vrpm install to make an easily reproduceable test which shows the misbehaviour you saw on ext3? 1120692863 M * Bertl also could you recompile your kernel with a simple change (no need to isntall it) 1120692903 M * Bertl in include/linux/vserver/xid.h change the first #warning to an #error 1120693042 M * baggins "config option missing" warning? 1120693169 M * Bertl yep, that one 1120693215 M * baggins done, make running 1120693243 M * baggins vserver fedora build -m apt-rpm -- -d fc3 1120693245 M * Bertl btw, you are right, we do the xid tagging in jfs, ext2/3, reiser and for nfs, but not xfs 1120693256 M * baggins [root@duch root]# ls -ld /vservers/fedora/var/spool/mail/ 1120693259 M * baggins drwxrwxr-x 2 root 4144959 6 Aug 12 2004 /vservers/fedora/var/spool/mail/ 1120693316 M * Bertl maybe you could use --debug and see what commands are executed to create that dir? 1120693349 M * Bertl I guess it will boil down to some more or less trivial sequence ... 1120693466 M * baggins /usr/sbin/vapt-get /etc/vservers/fedora -- install -y --reinstall filesystem 1120693707 M * baggins going deeper: 1120693711 M * baggins /usr/sbin/vnamespace --new -- /usr/lib64/util-vserver/vapt-get-worker /etc/vservers/fedora install -y --reinstall filesystem 1120693749 M * baggins it ends up as 1120693761 M * baggins /usr/lib64/util-vserver/vrpm-preload -Uvh filesystem_2.3.0-1_i386.rpm 1120693785 M * Bertl hmm, okay, let's try something like: 1120693878 M * Bertl chcontext --xid 256 rpm --root=/vserver/name -Uvh /path/to/filesystem_2.3.0-1_i386.rpm 1120693893 M * Bertl maybe with an additional -- before the rpm 1120694069 M * baggins drwxrwxr-x 2 root mail 6 Aug 12 2004 /vservers/fedora/var/spool/mail/ 1120694076 M * baggins looks good to me 1120694148 M * Bertl and you are 100% sure that you see those issues on ext3 too? 1120694229 M * baggins give me few minutes I run it on ext3 1120694240 M * Bertl k, np 1120694336 M * baggins positive, same on ext3 1120694360 M * Bertl and on ext2 it's fine?