1125964828 M * mugwump maybe reshed time is not quite right. but when the task is taken off active. 1125964892 M * mugwump do you know if the cache line issue happens every time you access it, or is it just happening because the ticks all occur at more or less the same time? 1125964939 M * mugwump ie, if we assume that tasks getting queued and dequeued from runqueues generally don't happen at the same time, then the conflict goes away if you batch up the token deductions to the end of each slice 1125964951 M * mugwump (runqueues on different CPUs) 1125964951 M * Bertl well, I was thinking of doing some borrowed time stuff with the process scheduling 1125964984 M * mugwump ah, so when you allocate a slice, take all the tokens for that slice from the bucket then 1125964995 M * mugwump then if it completes with tokens left, you can put them back 1125964998 M * Bertl i.e. the process withdraws the 'projected' amount of tokens from the TB, then runs for some time, and if unscheduled, puts them back 1125965012 M * Bertl yep, precisely 1125965068 M * Bertl that would even work at two levels, one for the context and the second per cpu 1125965093 M * mugwump how would that work? would you have one TB per cpu per context? 1125965101 M * Bertl yes 1125965161 M * mugwump won't you still have cacheline issues, unless each CPU's TBs are in seperate "pages" of memory? 1125965176 J * Nicoli ask@208.53.159.170 1125965198 M * Bertl mugwump: we would align the per cpu data at different cachelines 1125965214 M * Bertl (cachline is usually 32 - 128 bytes) 1125965229 M * Bertl welcome Nicoli! 1125965233 M * Nicoli thx 1125965235 M * mugwump ah, that's not so bad. and the correct value thereof is easily accessible? 1125965237 M * Greek0 Hollow: still awake? 1125965274 M * Bertl mugwump: yes, it is already used in per cpu variables, works on most archs except ia64 for now (which has different issues) 1125965328 M * Bertl Nicoli: atalkd? 1125965350 M * Nicoli eh? 1125965385 M * Bertl was just a shot in the dark :) 1125965403 M * Nicoli whatever that lang is, i don't speak it heh 1125965422 M * Bertl np, nevermind ... 1125965465 M * mugwump but ia64 is the best arch ever! even better than pa-risc 1125965472 M * Nicoli lol 1125965485 M * Bertl mugwump: yes, especially since version II :) 1125965487 P * hillct 1125965677 M * Bertl mugwump: what is your cpuhog command? 1125965686 M * mugwump I just use perl -e '1 while 1' 1125965742 M * mugwump the buckets seem to work until vsched is used. then it gets weird and even stopping/starting doesn't get them right again 1125965921 M * Bertl hmm .. did you enable vserver debugging yet? 1125965959 M * mugwump I thought I did, but sadly it doesn't seem to be set in /proc/config.gz 1125966570 M * mugwump joy! my modules are finished building, let's see if that worked 1125966575 A * mugwump reboots yet again 1125967488 M * mugwump ok, out to lunch ... back in the afternoon. 1125967500 M * mugwump I can't explain it, but that change I made certainly solves the symptoms 1125967538 J * click click@dsl-static-122-208.aal.tiscali.no 1125969581 Q * Aiken Remote host closed the connection 1125969589 J * Aiken ~james@tooax6-190.dialup.optusnet.com.au 1125971505 Q * SG_ Ping timeout: 480 seconds 1125972340 Q * jkl Quit: BitchX-1.1-final -- just do it. 1125972655 Q * litage Read error: Connection reset by peer 1125973377 J * jkl eric@c-71-56-237-229.hsd1.co.comcast.net 1125973402 M * jkl any hints on getting paralell printers working with cups in a vserver? 1125973418 M * jkl i'm getting "Parallel port busy; will retry in 30 seconds..." 1125973431 M * Bertl welcome jkl! 1125973432 M * jkl and -bash: lp0: Device or resource busy 1125973445 M * Bertl do you have a device for the parport inside the guest? 1125973447 M * jkl oh, whats up bertl! 1125973459 M * jkl yes, inside the guest i can see /dev/lp0 1125973479 M * Bertl maybe the host uses it too? via cups? 1125973490 J * litage ~nick@203.201.97.143 1125973497 M * jkl don't believe so, lemme check 1125973499 M * Bertl try with fuser /dev/lp0 on the host 1125973549 M * jkl on the host, i get no output 1125973590 M * jkl when typing that command 1125973605 M * Bertl hmm .. does it work on the host? cups I mean? 1125973623 M * jkl don't have it installed on the host 1125973646 M * jkl if i run that fuser command on the guest, i get a pid number 1125973683 M * Bertl which one? 1125973692 M * Bertl I mean, what process has that pid? 1125973697 M * jkl parallel:/dev/lp0 1 root Test Page 1 1125973719 M * jkl i may try and kill it and try the print again 1125973745 M * Bertl do you know that the partport works as expected? 1125973750 M * Bertl *parport 1125973769 M * jkl i'm not sure what you mean, how would i test that? 1125973844 M * Bertl well, it seems your assumption is that the 'failing' cups/parport issue is caused by linux-vserver, no? 1125973869 M * Bertl so I assume that you verified somehow that it works without linux-vserver, right? 1125973871 M * jkl i was hoping that was the problem, yes. 1125973886 M * jkl hehe, i will admit I am trying to cut some corners 1125973895 M * jkl ill see if i can get it to work on the host first 1125973909 M * Bertl probably a good start to narrow the issue down ... 1125973918 M * jkl absolutely 1125973981 M * jkl i just didnt want to have to setup cups on the host 1125974072 M * Bertl copy the config from the guest ... 1125974917 M * jkl ah, got cups printing test pages from the guest 1125974922 M * jkl now to get samba working 1125974933 M * jkl thanks for talking some sense into me, bertl :) 1125974998 M * Bertl you're welcome! what was the issue btw? 1125975278 M * jkl not quite sure exactly 1125975291 M * jkl i believe i tried to setup the printer with a PCL 6 driver, and tried a test page 1125975294 M * jkl that process froze 1125975312 M * Bertl k, so something cups specific ... 1125975313 M * jkl i reinstalled the printer without the PCL 6 driver, but it couldnt print because of that frozen process 1125975322 M * jkl yes it seems that wya 1125975324 M * jkl way even 1125976804 M * litage how do you clone a vserver? i can't find the docs on linux-vservers.org 1125976821 M * Bertl cp -va ? 1125976887 M * litage oh, that's it? hahah i thought there was an actual vserver program that did it 1125976923 M * Bertl well, you need to create the config for a new guest 1125976936 M * Bertl but that can be easily done with the -m skeleton build method 1125976962 M * Bertl so basically it's a) creating the config, and b) copying over the contents 1125977223 M * litage Bertl: so it'd be this?: vserver NAME build -m skeleton --hostname HOSTNAME --interface INTERFACE:IP_ADDRESS/PREFIX && cp -va /PATH/TO/CLONE/ /etc/vservers/.defaults/vdirbase/NAME/ 1125977301 M * Bertl modulo the pathes, yes 1125977333 M * litage "modulo the paths"? 1125977527 M * Bertl well, as the guest's dir will exist (/etc/vservers/.defaults/vdirbase/NAME/) it would probably copy /PATH/TO/CLONE into that, not the contents of /PATH/TO/CLONE 1125977604 M * litage ah gotcha. so just put a "*" after "/PATH/TO/CLONE/" and it'll copy the contents of CLONE/ into vdirbase/NAME/ 1125977634 Q * jkl Quit: BitchX-1.1-final -- just do it. 1125977645 M * Bertl or maybe '.' 1125977706 M * litage any difference? 1125977756 M * Bertl '*' will be resolved by the shell 1125977768 M * litage ah 1125977774 M * litage thanks, Bertl 1125977852 M * Bertl you're welcome! 1125978182 M * litage when i stop my vserver, it complains about modules.dep not existing, not being root. how can i fix this?: http://rafb.net/paste/results/lsTLo264.html 1125978206 M * litage s/existing, not/existing, and not/ 1125978342 M * Bertl disable the /etc/rc6.d/K89firewall.sh script 1125978358 M * Bertl (either remove the link, or unconfigure it) 1125978387 M * Bertl same goes basically for the network and umount stuff 1125978404 M * litage Bertl: shouldn't the vserver stop its firewall before going down? 1125978421 M * Bertl the vserver guest has no firewall, so nothing to stop there :) 1125978428 M * litage Bertl: i installed my own firewall 1125978447 M * Bertl fine, but the guest will not be allowed to modified iptables and such 1125978460 M * Bertl so it's a little pointless ... 1125978476 M * litage ah =P 1125978491 M * Bertl you can have firewalling on the host though ... 1125978495 M * litage yup 1125978533 M * litage Bertl: i didn't change anything concerning mounts though, except for the cd-rom in HOST:/etc/vservers/GUEST/fstab 1125978603 M * Bertl yes, you can ignore the errors, or clean up the scripts ... whatever you prefer 1125978681 M * litage Bertl: which script(s) can i clean up to remove the umounting messages? 1125978779 M * Bertl check the links in /etc/rc6.d/* 1125978789 M * Bertl one of them is probably doing the 'shutdown' 1125978794 M * litage k 1125978814 M * Bertl check for those messages 1125978814 M * Bertl Saving random seed...done. 1125978814 M * Bertl Deconfiguring network interfaces...done. 1125978814 M * Bertl Cleaning up ifupdown...done. 1125978814 M * Bertl Deactivating swap...umount: none: not found 1125978817 M * Bertl umount: /tmp: must be superuser to umount 1125978827 M * Bertl Unmounting local filesystems...u 1125978985 M * litage Bertl: do i even need the umountfs rc-script? 1125978999 M * Bertl no, not at all 1125979055 M * litage btw, if i've made this guest on boxA and then copy it to boxB (which has different hardware), will the guest work on boxB? 1125979196 M * Bertl yes, that is one of the advantages ... only requirement is a compatible architecture ... 1125979227 M * Bertl for example you can move an x86 guest to an x86_64 host (just requires ia32 support) 1125979240 M * litage awesome possum! 1125979301 M * litage Bertl: sorry to keep peppering you with questions, but i've got another :) you said that the host takes care of all hardware and network matters, so it's safe to remove /etc/init.d/networking from the guest? 1125979347 M * Bertl yes, it is 1125979402 M * litage thanks again! 1125979410 M * Bertl my pleasure ... 1125979665 M * litage Bertl: i've noticed that the reboot command in guest:/etc/init.d/reboot uses the -i option, which shuts down the network interfaces. i guess i'll remove that too =P 1125979704 M * Bertl good idea! 1125979831 M * Bertl okay, I'm off to bed now ... back in the evening ... 1125979849 M * Bertl have fun everyone! 1125979854 N * Bertl Bertl_zZ 1125980805 J * Adamsgr8 ~adam@12-208-125-171.client.insightBB.com 1125980876 P * Adamsgr8 1125981270 Q * Aiken Quit: Leaving 1125981553 J * Aiken ~james@tooax6-190.dialup.optusnet.com.au 1125982639 J * jpacheco ~justin@CPE0050180cd14b-CM0f0099806976.cpe.net.cable.rogers.com 1125982645 M * jpacheco hey guys! 1125982663 M * jpacheco question: is there a patch out yet that goes against the 2.6.13 kernel? 1125982754 M * mugwump http://vserver.13thfloor.at/Experimental/patch-2.6.13-vs2.0.1-pre2.diff.bz2 1125982795 M * jpacheco mugwump: thx, this site is so hard to navigate 1125982805 M * mugwump that's why we have irc :) 1125982832 M * jpacheco lol 1125982850 M * jpacheco mugwump: is this patch stable enough to use? 1125982895 M * mugwump that is the prerelease of the next stable patch ... no new features since 2.0 1125982910 M * mugwump if you want unstable, it is http://vserver.13thfloor.at/Experimental/patch-2.6.13-vs2.1.0-pre6.diff.bz2 1125982924 M * litage what might cause sudo to act strangely `sudo ls` returns: sudo: unable to lookup camel via gethostbyname() 1125983054 M * mugwump does perl -le 'print((gethostbyname("camel"))[0])' work? 1125983122 J * hillct ~hillct@client200-5.dsl.intrex.net 1125983149 M * litage mugwump: returns nothing 1125983193 M * mugwump well, there you go.. you can't resolve your hostname as returned by uname -n 1125983211 M * mugwump check your resolver config, hosts file, etc 1125983222 A * mugwump & # out partying 1125983222 M * litage mugwump: `uname -n` returns "camel" though 1125983232 M * jpacheco mugwump: have you tried 2.6.13 yet? 1125983232 M * mugwump yes, but that's not in your /etc/hosts 1125983238 M * mugwump jpacheco: not yet 1125983260 M * jpacheco damn, im trying to find someone that has inotify working 1125983335 M * litage hrm, i don't even have /etc/hosts =/ 1125983404 M * litage should a vserver/guest's /etc/hosts file be in the vserver/guest, or on the host in /etc/vservers/GUEST/ ? 1125983447 M * mugwump inside the guest should be "complete"... it can't contact the outside! 1125983546 M * hillct was there a behavior change in the way sync-timeout defaults to between util-vserver 206 and 208? 1125983576 M * hillct I'm suddenly getting timeouts waiting for KillAll in guest shutdown 1125983670 M * litage mugwump: gotcha. i thought that since the host takes care of the networking, it might also take care of the hosts file 1125987411 M * Hollow morning 1125987650 M * dddd44 how do i know which font my web browser displaying by looking this code anyone?? body { 1125987650 M * dddd44 background-color: white; 1125987650 M * dddd44 color: black; 1125987650 M * dddd44 font-family: 'Trebuchet MS', Verdana, Tahoma, Helvetica, Geneva, sans-serif; 1125987650 M * dddd44 } 1125987652 J * eyck eyck@81.219.64.71 1125987943 M * eyck hello 1125987988 Q * jpacheco Quit: BitchX-1.1-final -- just do it. 1125988037 P * hillct 1125988780 M * Hollow dddd44: i suggest looking at "font-family" 1125988960 M * eyck vserver depends on fonts now? 1125989138 M * Hollow not that i know of ;) 1125989160 M * Hollow he just pasted a css block and wanted to know want font is used 1125989175 M * Hollow well, the easiest thing is to look at the font-family attribute ;) 1125989190 M * Hollow but not sure if that even the right channel to ask such things :P 1125989200 M * Hollow i no verb 1125989201 M * Hollow ;) 1125989348 M * Greek0 re 1125989524 M * eyck sure it is... 1125989528 M * eyck we know everything. 1125989704 N * _pg|cereal cereal 1125991763 J * _Hunger Hunger.hu@Hunger.hu 1125991864 Q * Hunger Ping timeout: 480 seconds 1125993516 N * _Hunger Hunger 1125993812 Q * cryo Ping timeout: 480 seconds 1125994682 M * Greek0 wow. 93 lines of swig and I've got libvserver python bindings :) 1125994806 J * kas_3 ~dhb55@tor-irc.dnsbl.oftc.net 1125994996 J * weasel weasel@weasel.noc.oftc.net 1125995024 J * Adamsgr8 ~adam@12-208-125-171.client.insightBB.com 1125995129 Q * Adamsgr8 Quit: 1125995220 Q * dddd44 Ping timeout: 480 seconds 1125995295 M * SiD3WiNDR python *shudder* 1125995552 M * Greek0 I'm pretty convinced that python tools would be quite a bit easier to maintain then c/bash 1125995562 M * Greek0 especially since I saw how util-vserver looks today 1125995585 M * eyck I'm pretty convinced that c/sh tools would be quite a bit easier to maintain then python 1125995625 M * SiD3WiNDR eyck++ 1125995642 M * eyck SiD3WiNDR: how about now - perl would be the best? ;) 1125995650 M * SiD3WiNDR eyck-- 1125995651 M * SiD3WiNDR :p 1125995688 M * Greek0 all that c/sh seems to be is the least common denominator.. 1125995723 M * Greek0 IMHO scripting all this not too easy/obvious stuff in sh isn't so great 1125995751 J * Tenchi-san ask@208.53.159.170 1125995753 M * eyck personally I would prefer perl, but I'm ready to admit that project on the whole could be better of with C. 1125995760 M * eyck off 1125995762 M * eyck even. 1125995784 Q * Nicoli Read error: Connection reset by peer 1125995838 M * Greek0 I think that doing it in sh sucks because sh just isn't very good at stuff of this scale and complexity. 1125995849 M * Greek0 and scripting everything in c seems a little over-complicated to me. 1125995861 M * Greek0 that's why I think a scripting language would be nice for it 1125995870 M * eyck hmm 1125995875 M * eyck sh is a scripting language 1125995889 M * Greek0 yep, just quite a limited one 1125995903 M * eyck really? where exactly? 1125995942 M * Greek0 interacting with running programs for example? 1125995971 M * eyck so, you want classical glue language? 1125995976 M * eyck perl and rexx come to mind. 1125995977 M * Greek0 I find that sh gets pretty complex/unobvious 1125995985 M * Greek0 yep. python comes to my mind :) 1125996003 M * SiD3WiNDR sick mind. 1125996004 M * SiD3WiNDR >:) 1125996171 M * eyck well, google thinks those are glue languagues: 1125996173 M * eyck perl 1125996174 M * eyck ruby 1125996178 M * eyck (???) 1125996182 M * eyck sh 1125996211 M * eyck http://wiki.linuxquestions.org/wiki/Glue_language&e=747 1125996216 M * SiD3WiNDR tcl! 1125996234 M * SiD3WiNDR (There is currently no text in this page) 1125996236 M * eyck correct. tcl was well-known glue lang 1125996238 M * SiD3WiNDR only ads 1125996271 M * eyck interesting, and this: http://www.google.pl/url?sa=U&start=3&q=http://wiki.linuxquestions.org/wiki/Glue_language&e=747 ? 1125996326 M * SiD3WiNDR yes, so you meant http://wiki.linuxquestions.org/wiki/Glue_language 1125996335 M * SiD3WiNDR :) 1125996348 M * eyck hmm, yeah, well, google knows best. 1125996363 M * Greek0 oh, I'm not thinking about such a glue language that the definition there says 1125996390 M * eyck yeah, you're basically thinking like this 'python, python, python....python' 1125996399 M * Greek0 I'd like to use the libvserver python bindings I just created. 1125996417 M * eyck yeah :) that's exactly what I mean 1125996419 M * SiD3WiNDR :p 1125996429 M * Greek0 eyck: ok. you can do the same thing in perl too 1125996429 M * SiD3WiNDR ah, so you're looking for an excuse for the time you just wasted on making them? :p 1125996441 M * Greek0 the swig thing can just as well generate perl bindings 1125996453 M * Greek0 or tcl or ruby or guile or whatever other language you want 1125996473 M * Greek0 and if you like to code something up in tcl, well, as nike told us: just do it 1125996486 J * prae ~prae@gut75-1-81-57-27-189.fbx.proxad.net 1125996526 M * Hollow python.. *chuckle* 1125996575 M * Hollow python is pure wishy washy 1125996582 M * Hollow just like php 1125996584 M * Hollow bah. 1125996611 M * eyck Hollow++ 1125996639 M * Hollow my utils'll never see any byte of python code :P 1125996661 M * Hollow but bindings are great anyway 1125996663 M * Greek0 I haven't said that everyone has to use python bindings, damnit. I just created a file that spits out bindings for some languages 1125996672 M * Hollow Greek0: mind uploading the swig source? 1125996692 M * Hollow i'm more interested in the ruby bindings ;) 1125996710 M * Greek0 Hollow: also nice, I don't know ruby all too much, but it sounds quite good too 1125996729 M * Hollow yeah, i like ruby, it's OOP (and it really is) 1125996745 M * Greek0 I'm not so python-zealotish as I sound occasionally. I just don't think c/sh is the best solution for the tools 1125996751 M * eyck Greek0: well, even google agrees that ruby is one of well-known glue languages. 1125996761 M * Hollow Greek0++ punt bash :) 1125996764 M * Greek0 eyck: ok. hail to google 1125996774 M * eyck yeah, ruby's OO is great :) 2.times ...sth.. :) rock. 1125996782 M * Hollow yep :) 1125996811 M * Hollow you can write the complexest things with so few lines of code 1125996895 M * Greek0 http://greek0.net/~greek0/div/swig.tgz 1125996923 M * Hollow thx 1125996944 M * Greek0 there are 2 files inside, vserver.i, the main file, and vserver-nx_addr.i, which wraps the arrays in nx_addr, so they can be accessed via lists 1125996951 M * Greek0 the latter is quite python specific 1125996966 M * Greek0 donno if this has to be done for ruby too or if it works out of the box there 1125996972 M * Greek0 (doubt it actually) 1125997043 M * Hollow i.e. swig does all binding automagically if you only iclude the header? o.o 1125997070 M * Greek0 more or less. arrays-in-structures don't really work 1125997084 M * Hollow rocks anyway ;) 1125997087 M * Greek0 and you need some typedefs so you can actually use the integer datatypes from your language 1125997103 M * Greek0 the most complex part was getting that array-stuff to work 1125997167 M * Greek0 IME we don't actually need more tools in c. I'd just use the libvserver bindings directly in $SCRIPTING_LANG, and write all the tools there 1125997217 M * Greek0 since you don't have to fork to execute some commands, and having each command as a function in your language gives you more flexibility IMHO 1125997219 M * Hollow Greek0: basically that's my idea too 1125997237 M * VooDooMaster good mornig :) ... where can I find some docs about how to "create" interfaces in a vserver (using host ip, own ip, interface alias, ...) ? 1125997248 M * Greek0 Hollow: so you plan to use ruby for that? 1125997252 M * Hollow VooDooMaster: http://www.nongnu.org/util-vserver/doc/conf/configuration.html 1125997279 M * VooDooMaster Hollow: Thanks ... 1125997294 M * Hollow Greek0: i'd like to, but i'm not sure if the dependency on ruby would be that good, as it's not that spread, is it? 1125997335 M * Greek0 hmm. I'm surprised to see that I have it installed 1125997340 M * Hollow heh 1125997372 M * Greek0 the most widely used lang is obviously perl, but at least for me that's not an option 1125997400 A * Hollow nods 1125997482 M * Greek0 but there are py2exe and rubyscript2exe 1125997495 M * Hollow Greek0: the nx_* functions are more or less untested/unused anyway 1125997526 M * Hollow at least i think the concept of those functions has to be rethought 1125997543 M * Greek0 hmm, so we could just use vserver.i to let ruby stuff drop out of swig, and start hacking :) 1125997551 M * Hollow yay! :) 1125997660 M * Greek0 hmm 6M libruby 1125997725 M * Hollow also not sure if vserver-util will be kept in the tree... this could also be implemented using $SCRIPT_LANG 1125997732 M * Hollow libvserver-util 1125997753 M * Hollow was merely hacking around ;) 1125997818 M * Greek0 mm, do you actually use libvserver-util in your current c utils? 1125997823 M * Hollow yup 1125997850 M * Hollow nearly every command line option that forms a list or hash is parsed using the list functions in libvserver-util 1125997870 M * Greek0 what is the addr to the svn of vserver-util? 1125997888 M * Hollow http://dev.croup.de/repos/vserver-util 1125997893 M * Hollow + /trunk ;) 1125997906 M * Greek0 ah, so it's not really backend-ish, just some misc tools that probably wouldn't be needed in dynamic languages anyway? 1125997916 M * Hollow yup 1125997997 M * Greek0 svn co http://dev.croup.de/repos/vserver-util doesn't work 1125998012 M * Hollow ar... it's vserver_utilS 1125998018 M * Hollow add s 1125998027 M * Greek0 ah :) 1125998074 M * Greek0 do you have ruby bindings already? 1125998095 M * Greek0 i.e. does it work just running it on vserver.i (with the vserver-nx_addr.i line removed)? 1125998104 M * Hollow no, still browsing through the swig doc ;) 1125998134 M * Hollow but lemme try 1125998198 M * Greek0 Vudumen: could you put the source for your .deb's online too please? 1125998264 M * Hollow have to merge swig first 1125998270 M * Greek0 swig -python lib/vserver.i; gcc -Wall -ggdb --shared lib/vserver_wrap.c -I/usr/include/python2.4/ -o lib/_vserver.so lib/.libs/libvserver.so.1 1125998298 M * Greek0 ^-- that's what I used for the python stuff. obviously from the root of the libvserver source tree, vserver.i was placed in lib 1125998384 M * Hollow hm, have to merge ruby too *g* 1125998434 M * Greek0 lol 1125998636 A * Greek0 's just reading Why's (poignant) guide to ruby 1125998808 A * Hollow 's just compiling swig 1125998810 M * Hollow ;) 1125998883 M * litage when i run `hostname NAME`, the guest's hostname is changed, but why isn't the new name put into /etc/hostname ? 1125998924 M * Greek0 litage: because the vserver stuff doesn't just change stuff on your filesystem 1125998957 M * Greek0 you can change /etc/hostname easily from the host, before starting the vserver 1125998961 M * litage Greek0: sorry, i meant "when i run `hostname NAME` within the guest, ..." 1125999044 M * Greek0 litage: because hostname(1) just changes the hostname of the running system. 1125999061 M * Greek0 /etc/hostname is used by the init scripts to automatically set the hostname at boot-time 1125999065 M * litage ah i see 1125999066 M * litage thanks 1125999174 M * Greek0 np 1125999278 M * litage i just changed the hostname with `hostname VSHOST && echo VSHOST >/etc/hostname` and rebooted the box, but `hostname` still returns the old name 1125999328 M * Greek0 litage: $ cat /etc/hostname 1125999346 M * Hollow wh00t 1125999359 M * litage Greek0: that outputs VSHOST 1125999359 M * litage Greek0: but `hostname` outputs the old name 1125999381 M * Greek0 did you do that in a vserver or on your main machine? 1125999399 M * litage in a vserver 1125999420 M * Greek0 ah, IIRC vservers can't set the system-wide hostname (per default) 1125999428 M * Greek0 you want the /proc virtualisation 1125999429 M * Hollow Greek0: http://home.xnull.de/misc/swig-ruby.log 1125999446 M * Hollow so.. let's see.. ;) 1125999451 M * Hollow install ruby first 1125999453 M * Hollow maybe.. 1125999454 M * Hollow *g* 1125999461 M * Greek0 Hollow: you're missing -I to the ruby include dir 1125999463 M * litage Greek0: ah, so once you've created a vserver, you can't change the hostname? 1125999470 M * Greek0 or you're missing ruby as a whole ;) 1125999479 M * Hollow seems so, yes ;) 1125999484 M * Hollow BAH! 1125999489 M * Hollow ruby needs gdbm 1125999497 M * Hollow i just got rid of that 1125999531 M * Greek0 litage: you can. look at http://www.nongnu.org/util-vserver/doc/conf/configuration.html 1125999534 M * Greek0 and search for hostname 1125999541 M * litage thanks 1125999554 A * Greek0 whistles something about debian and automatic dependency tracking and stuff 1125999570 M * Hollow ? 1125999601 M * Greek0 (versioned) dependency tracking of packagesß 1125999603 M * Greek0 -' 1125999609 M * Greek0 err. -ß 1125999668 M * Hollow ehm, gentoo does it too ;) 1125999675 M * Greek0 but? 1125999685 M * Hollow but the ebuild lists gdbm as a dep... 1125999691 M * Hollow and i don't want gdbm ;) 1125999712 M * Hollow if i want to use a database i use sql, not dbm 1125999722 M * Greek0 ic 1125999736 M * Hollow so i got rid of that stuff 1125999750 M * Hollow and i don't like ebuilds forcing me on gdbm if i won't need it (ruby doesn't need it) 1126000034 M * Greek0 hmm. does anyone know if Vudumen has the source for his .deb packages online somewhere? 1126000202 J * Neubix ~brian@p54B044C0.dip.t-dialin.net 1126000207 M * Neubix moin ... 1126000250 M * Neubix someone know how to install SuSE Linux as guest without CU-drom (over internet) ?? 1126000548 M * SiD3WiNDR cudrom? =) 1126000552 M * SiD3WiNDR and no, sorry :/ 1126000558 M * SiD3WiNDR I guess they have a floppy set to boot from? 1126000564 M * Neubix cdrom !! :-) 1126000595 M * Neubix boot from ?? guest ... !! not vserver host 1126000625 M * Neubix The server is 30 Km from me .. and no floppy 1126000631 M * wibble hmm 1126000641 M * wibble vdlimit quotas don't stick on reboot 1126000656 M * Hollow wibble: vsched as well 1126000682 M * wibble hmm, any reason for this? 1126000695 A * Hollow shrugs 1126000723 M * Hollow vsched settings get lost when the context dies, but dunno why vdlimit won't keep it's settings 1126000796 M * wibble that's a bit of a bummer 1126000803 M * Neubix or is there a Suse image to download ? 1126000863 M * Greek0 Hollow: why did you actually remove the dietlibc dependency from libvserver? 1126000965 M * Hollow it currently doesn't build with dietlibc 1126000992 M * Greek0 haven't tried it, why? 1126001023 M * Greek0 I mean I've heared quite some times that glibc had some problems when being used for the vserver tools 1126001063 M * Greek0 e.g. that it checks /etc/resolv.conf (which can already be the guest file, not the one of the host) 1126001067 M * wibble hmm 1126001069 M * Hollow the only reason (mentioned in the util-vserver package) is the nss functions ins glibc, and we don't use them in libvserver anyway 1126001075 M * wibble how do I run something once the guest has started? 1126001086 M * wibble i.e. set disk quota limit :) 1126001136 M * Greek0 wibble: there are hooks (check the flower page), alternatively in the guest init scripts 1126001148 M * Greek0 depends on who wants to get something run, the host admin or the guest admin 1126001167 M * wibble the host admin 1126001267 M * Greek0 hmm. am I stoned now? I thought there was some hooking stuff in util-vserver 1126001321 M * Greek0 ah, there are. /etc/vservers/vserver-name/scripts 1126001395 M * Greek0 pre-start might be useful, but I don't know what's that current root at that point of time (i.e. if rbind /vservers/.. /; chroot /; has already taken place 1126001436 M * Hollow bene@zeus [0] build $ irb 1126001437 M * Hollow irb(main):001:0> require 'vserver' 1126001437 M * Hollow LoadError: ./vserver.so: undefined symbol: vx_migrate - ./vserver.so 1126001437 M * Hollow from ./vserver.so 1126001438 M * Hollow from (irb):1 1126001446 M * Hollow :( 1126001541 M * Greek0 swig -ruby vserver.i; gcc -Wall -ggdb --shared vserver_wrap.c -I /usr/lib/ruby/1.8/i486-linux/ -o vserver.so /usr/lib/libvserver.so.1 1126001579 M * Greek0 I think you need that /usr/lib/libvserver* on the command line to get the NEEDED entry in the vserver.so, so libvserver.so.1 is loaded by the dynamic linker 1126001595 M * Hollow yeah, i did that 1126001607 M * Hollow bene@zeus [0] build $ gcc -c ../lib/vserver_wrap.c -I/usr/lib/ruby/1.8/i686-linux 1126001607 M * Hollow bene@zeus [0] build $ gcc -shared /usr/lib/libvserver.so vserver_wrap.o -o vserver.so 1126001637 M * Greek0 I gotta build custom .debs or something. Vudumen's are linked against dietlibc, so my builded stuff doesn't work against them 1126001714 M * Greek0 works for me now 1126001750 M * Hollow d'oh 1126001765 M * Hollow yeah... 1126001770 M * Hollow was some opther error in my working copy 1126001776 M * Greek0 fine 1126001909 A * Greek0 rolls with his eyes "elsif" ?! 1126001971 M * wibble ok quick script written 1126001996 M * eyck elsif rocks. 1126002039 M * Greek0 hell, elsif. what were they smoking when they came up with this 1126002063 M * eyck not pythongs obviously 1126002065 M * Greek0 I mean, else if, elif, yes.. but I've never ever seen elsif before 1126002098 M * eyck well, this says loads about your qualifications as a programmer. 1126002117 M * Greek0 what language uses elsif? 1126002134 M * Greek0 and btw, I didn't comment on your ability as programmer when you were bashing python for no apparent reason 1126002290 M * Greek0 ah ok. perl seems to use it 1126002300 M * Greek0 and vhdl 1126002488 M * Hollow Greek0: semi works 1126002500 M * Hollow it can load the module 1126002507 M * Hollow irb(main):008:0> puts Vserver.vx_get_task_xid(0) 1126002507 M * Hollow TypeError: wrong argument type Fixnum (expected Data) 1126002524 M * Hollow so, there are some type errors 1126002555 M * Greek0 ok, I'm still stuck in the "getting some basic understanding of ruby" phas 1126002556 M * Greek0 e 1126002606 M * Hollow :) 1126002609 M * Greek0 how can I get a list of the contents of the module? 1126002620 M * Greek0 and is there some tab completition for irb? 1126002638 M * litage is it okay to chmod 700 /etc/vservers/.defaults/vdirbase/NAME ? 1126002650 M * Greek0 print Vserver.vx_get_task_xid(0) seems to work for me 1126002655 M * Hollow shrug@both ;) 1126002661 M * Hollow but the latter i don't think so 1126002678 M * Greek0 returns -1, which is probably right, since I'm testing this on my laptop without vserver patches 1126002722 M * Hollow hm 1126002734 M * Greek0 as does puts 1126002741 M * Greek0 swig version? 1126002759 M * Hollow 1.3.21 1126002764 M * Greek0 1.3.24-1 here 1126002826 M * Greek0 in the vserver.i file anything else done then commenting out the %include lib/vserver-nx_a....i line? 1126002846 M * Hollow yup, i just included the header 1126002851 M * Greek0 especially you really need all the {u,}int* typedefs there 1126002855 J * brian_ ~brian@p54B042DD.dip.t-dialin.net 1126002855 M * Hollow maybe add back these typedefs 1126002860 M * Greek0 :P 1126002863 M * Hollow *g* 1126002893 M * Greek0 they are so that swig knows that uint32_t* aren't some opaque data types or structures, but really ints 1126002916 M * Greek0 I know it sucks a bit to hardcode the int32 -> unsigned long int mappings there.. 1126002929 M * Greek0 donno if there is a better alternative 1126002985 M * Hollow yeah :) 1126003001 M * Hollow works 1126003262 M * Greek0 could you rewrite some simple program from your vserver-util stuff with those bindings so I can have a look at it (to get a better feeling for ruby) 1126003294 M * Hollow Greek0: http://phpfi.com/77284 1126003298 M * Hollow some first step ;) 1126003315 Q * Neubix Ping timeout: 480 seconds 1126003406 M * Greek0 IMHO everything should work fine except nx_addr. 1126003419 M * Hollow yup 1126003461 Q * brian_ Quit: Verlassend 1126003466 M * Greek0 vx_vhi_name should work well, even with bounds checking of the string 1126003488 M * Greek0 it at least did in python (wow! I said python again. I must be truely crazy for python!!!!) 1126003497 M * Hollow python sucks 1126003499 M * Hollow :P 1126003529 M * Greek0 and I guess you saw the warnings about the const char *'s .. that they might leak memory 1126003548 M * Hollow yup 1126003581 M * Hollow i roughly tested libvserver with valgrind, so dunno.. 1126003583 M * Greek0 I just thought we'd worry about that when everything turns out to work.. shouldn't be too much of a problem to fix within swig 1126003593 A * Hollow nods 1126003603 M * litage when i ssh into a guest that i just created, i end up ssh'ing into the host. how can i fix this? 1126003628 M * Hollow litage: let sshd on the host only listen to the hosts ip and not 0.0.0.0 1126003638 M * litage Hollow: ah right. my bad. thanks 1126003798 M * Hollow Greek0: main question is if we want to provide a full-blown ruby/whatever module (exception handling e.g.), or just the simple wrapper around the syscalls 1126003919 M * Greek0 I guess if we really implement all the tools on top of those bindings, we'll automatically get the former, not? 1126004011 M * Hollow seems so 1126004864 M * nokoya hello ppl 1126004874 M * nokoya I got problem with ssh 1126004887 M * nokoya I tried connect thru ssh 1126004897 M * nokoya I saw the screen ask me about username 1126004905 M * nokoya and then I type it and press enter 1126004920 M * nokoya but after that I didn't see the screen ask me the password 1126004930 M * Greek0 are you sure you have connected to the sshd in the vserver, not the one on the host? 1126004931 M * nokoya I can login 2 days ago but not tomorrow 1126004942 M * nokoya yes 1126004948 M * nokoya I checked the sshd_config 1126004951 M * nokoya no problem too 1126004965 M * Greek0 strange. did you do upgrades somewhere? 1126004971 M * nokoya no 1126004987 M * nokoya I didn't touch anything in 1 week b4 1126004991 M * Greek0 on my debian boxes I hit some strange ssh problem some weeks ago, upgrading to the latest versions helped 1126005038 M * nokoya is it sshd problem ? 1126005039 N * BobR_oO BobR 1126005039 M * nokoya seems wierd 1126005068 M * Greek0 mm. weird indeed. no idea what might be wrong.. 1126005092 M * Greek0 I'd just suggest upgrading both ends of the ssh connection to the latest version available and retrying 1126005315 M * VooDooMaster hi! which (simple/command line) tool do you guys recommend for monitoring network devices (bandwidth usage, transfer volume, ..)? it should produce a "nice" but not too fancy text or html/graphical output. so I can "see" what the root-server and the single vservers network usage is 1126005352 M * litage i chmod'd /etc/vserver/.default/vdirbase/GUEST/ 700. when starting the guest, why would the nullmailer daemon fail to start?: Unable to start /usr/sbin/nullmailer-send: Permission denied 1126005413 M * Greek0 litage: because nullmailer-send wants change to some unprivileged user, and then can't access and/or execute anything anymore? 1126005478 N * BobR BobR_afk 1126005525 M * litage Greek0: when /etc/vserver/.default/vdirbase/GUEST/ was chmodd 755, this error didn't occur. why do the permissions on that dir affect what runs inside the guest? 1126005557 M * Greek0 that's the new / of the guest, not? 1126005570 M * litage yes 1126005587 M * Greek0 and how should nullmailer-send open some file inside the guest namespace, if it doesn't even have access rights to / 1126005617 M * Greek0 since nullmailer-send probably setuid()'s to some unprivileged user for improved security 1126005629 M * litage by that time, isn't nullmailer-send running within the guest and not the host though? 1126005663 M * Greek0 and remember that when you access /path/to/some/file, not only the parms for file are checked, but also those for /path, /path/to and /path/to/some 1126005693 M * Greek0 where do you have nullmailer-send running? inside the guest or outside? 1126005699 M * litage Greek0: inside the guest 1126005718 M * Greek0 ok. so nullmailer-send su's to some user 1126005755 M * Greek0 then it wants to open /var/mail/something. permission check: read/readdir permission for /, for /var and for /mail 1126005774 M * Greek0 permission check fails right at the beginning, because unprivileged users can't read anything inside / 1126005786 M * Greek0 since that's chmod 700 root:root (or some other user) 1126005813 M * Greek0 if you don't want unprivileged users on the host to spy into the guests you could do something like this: 1126005882 M * Greek0 /vservers/ <-- normal vserver dir, mode 755. /vservers/private : base dir for private vservers, 700. and /vservers/private/your-vserver: 755 1126005901 M * Hollow why not just chmod 000 /vservers/ ? 1126005925 M * Greek0 Hollow: possible too, unless you want to mix private and non-private vservers of course 1126005973 M * litage Greek0: if i want all vservers to be private, is it okay to do this?: chmod 700 /etc/vservers/.defaults/vdirbase/ && chmod 700 /etc/vservers/.defaults/vdirbase/* 1126005992 M * litage "private" as in inaccessible by unprivileged host users 1126006002 M * Greek0 litage: no, you need 755 on vdirbase/* 1126006021 M * Greek0 otherwise normal users in the guest don't have permissions to access / 1126006024 M * Greek0 (in the guest) 1126006046 M * Greek0 chmod 700 /etc/vservers/.defaults/vdirbase/ && chmod 755 /etc/vservers/.defaults/vdirbase/* 1126006049 M * Greek0 ^-- IMHO 1126006090 M * litage Greek0: hahahah oops that 2nd "700" that i put in my previous msg was meant to be 755, as putting "700" would be ignoring all of what you just said 1126006107 M * Greek0 litage: hehe yes :) 1126006118 M * litage sorry about that typo; getting late here 1126006134 M * litage thanks for this tip, Greek0 1126006140 M * Greek0 np 1126006284 M * litage Greek0: since you've been so helpful, would you care to name my newest vserver? =P i'm naming them after animls (eg: file server == camel, log server == elephant). what would you call an APT Repository server? 1126006293 M * litage s/animls/animals/ 1126006298 J * Neubix ~brian@p54B042DD.dip.t-dialin.net 1126006315 M * Greek0 hmm. 1126006328 M * litage mail == albatross 1126006334 M * litage i like that one. heh 1126006344 M * Greek0 I liked the linux 2.6.0 codename: stoned beaver.. 1126006354 M * litage hahah 1126006364 M * litage that's 2 words though 1126006367 M * Greek0 and when I think of beavers they are a bit fat, just like the debian archive 1126006388 M * Greek0 just beaver then, if you like :) 1126006394 M * Hollow heh 1126006407 M * litage sounds good. and i'm canadian, too, so it's even better! 1126006413 A * litage : zing! 1126006415 M * Greek0 thanks for the honor :) 1126006425 M * litage thanks for the help 1126006459 M * litage Greek0: you KNEW there'd be perks to joining this project, didn't you! 1126006467 M * Hollow i took the greek approach :P 1126006518 M * litage Hollow: what's the greek approach? 1126006547 M * Greek0 great. no idea what "perk" means, and dict.leo.org down :-/ 1126006550 M * Hollow my hosts are named like greek gods 1126006557 M * Greek0 lol 1126006559 M * Hollow dict.leo.org is just danm slow 1126006566 M * Hollow but it works 1126006594 M * Greek0 mine like characters from the greek mythology: iphigenie, orest, phaon, OpenWRT 1126006606 M * Hollow Greek0: yeah, that mostly are gods 1126006625 M * Hollow http://home.xnull.de/misc/hosts ;) 1126006629 M * weasel Greek0: then use dict. doesn't give you a translation but english-english is better for you anyway :) 1126006638 M * Hollow heh 1126006657 M * litage at the last company i worked at, they used car parts: engine, trunk, wheel, gear, stick, dash, piston, exhaust, etc. 1126006678 M * Greek0 weasel: ? 1126006682 M * Hollow heh, i worked at a lotr company.. ;) 1126006697 M * Hollow smeagol etc.. 1126006704 M * Greek0 Hollow: afaik those orest, phaon, iphigenia were half-gods at best 1126006707 M * Hollow it's so much fun to name computers :P 1126006724 M * weasel Greek0: 'dict foo' on your favorite unix box 1126006744 M * litage Hollow: does this give both ip addresses the specified hostname?: apollon home.xnull.de 82.135.28.84/192.168.0.52 1126006757 M * Greek0 weasel: command not found ;) 1126006775 M * Hollow litage: these are local boxes which share one public ip through NAT 1126006784 M * Greek0 I have my firefox shortcut "dict ", that's been enough for a while 1126006797 M * litage Hollow: is that valid syntax for /etc/hosts? 1126006800 M * Hollow no 1126006825 M * litage oh okay. i thought you'd just cat'd and cut'd /etc/hosts 1126006835 M * Hollow no, just a list of my boxes ;) 1126006923 M * litage when was the linux-vservers project begun? 1126007097 M * eyck long long time ago, in a galaxy far far away 1126007111 M * litage heh really 1126007122 M * Hollow litage: 2001-10-02 12:08:28 was the first changelog 1126007139 M * litage i only found out about it ~2 weeks ago, but holy cow, i can see myself using it for years. it's sofa king useful 1126007170 Q * Hollow Remote host closed the connection 1126007227 M * eyck yeah, you said it mon 1126007247 J * Hollow ~Hollow@home.xnull.de 1126007730 M * Hollow skype sucks 1126007949 M * eyck really? 1126007974 M * litage thanks for your help guys, i really appreciate it! 1126008018 M * eyck no problem! 1126008125 M * eyck or maybe: 1126008129 M * eyck you're welcome! 1126008558 J * micah_ micah@micha.hampshire.edu 1126008977 Q * micah Ping timeout: 480 seconds 1126009002 J * Blissex ~Blissex@82-69-39-138.dsl.in-addr.zen.co.uk 1126009072 Q * matta Ping timeout: 480 seconds 1126010200 Q * douglas Ping timeout: 480 seconds 1126010768 N * BobR_afk BobR 1126010861 J * dddd44 ~dhb55@tor-irc.dnsbl.oftc.net 1126011155 Q * kas_3 Ping timeout: 480 seconds 1126011178 J * kas_3 ~dhb55@tor-irc.dnsbl.oftc.net 1126011321 J * ritch ~ritch@adsl-065-007-136-239.sip.asm.bellsouth.net 1126011354 P * ritch 1126011370 Q * dddd44 Ping timeout: 480 seconds 1126011709 J * keyser_soze ~cimarron@host163.201-252-24.telecom.net.ar 1126011740 Q * keyser_soze Quit: 1126012035 J * ydupont ~dupont-y@tomintoul.cri.univ-nantes.fr 1126012277 T * HostingGeek http://linux-vserver.org/ | latest stable 2.0, 2.0.1-pre2, 1.2.10, 1.2.11-rc1, devel 2.1.0-pre5 | He who retarts the topic is a retart for a lifetime not only a minute 1126012292 M * HostingGeek I didn't retart the topic! 1126012298 M * HostingGeek the guy before me did 1126012335 M * wibble right 1126012347 M * wibble i have a script to sort out the vdlimit quotas on reboot 1126012352 M * wibble very crufty but werks 1126012362 T * HostingGeek http://linux-vserver.org/ | latest stable 2.0, 2.0.1-pre2, 1.2.10, 1.2.11-rc1, devel 2.1.0-pre5 | Dear Mr Foo, we can not answer your question within 1 minute it is simply impossible... please give us at least 1 hour 1126013393 M * VooDooMaster Hi! does ccache speed-up emerges in vservers? what di I have to "mount" into th vservers, so that they can take advantage of the cache? 1126013427 M * eyck I would say, you have to mount ccache cache. 1126013610 Q * click Remote host closed the connection 1126013614 J * click click@dsl-static-122-208.aal.tiscali.no 1126013751 M * VooDooMaster eyck: I thought so ... ;) ... nothing more? ... 1126013882 M * VooDooMaster can someone explain me how to setup the network in vservers ... if possible ALL versions ... the docs I found are not really good :| ... I need bridged networking with same IF an dIP as host, IF-alias ith same IP as host, and IF-Alias with different IP as host. 1126013960 Q * VooDooMaster Quit: Nettalk6 der Freeware IRC-Client 1126013963 J * VooDooMaster VooDoo@topas.informatik.uni-ulm.de 1126014101 Q * Neubix Quit: Verlassend 1126014254 M * Greek0 hmm. networking isn't so hard. just stuff your config into /etc/vservers//interfaces 1126014301 M * Greek0 it's also quite intuitive, just not for most people (i.e. for everyone who doesn't know how the linux kernel handles the networking stuff ;) 1126014400 M * eyck Greek0: didn't he said "ALL versions" ? 1126014486 M * VooDooMaster eyck: or all variations ... 1126014490 M * Greek0 yep. he said all versions 1126014522 M * VooDooMaster Greek0: I looked at the flower page ... but that didnt't help :| ... 1126014547 M * Greek0 VooDooMaster: do you have problems with getting an interface setup inside the vserver in the first place, or in making it work as expected? 1126014553 M * VooDooMaster Greek0: How do I get hidden interfaces? ifconfig should show no interface but networking should be possimle in the vserver 1126014573 M * VooDooMaster Greek0: Making it work as I wish. 1126014582 M * Greek0 VooDooMaster: well, that's not possible IMHO 1126014595 M * Greek0 or better AFAIK 1126014597 M * VooDooMaster the new vserver-new script ist really cool - but I want to fine-tune my networking 1126014613 M * Greek0 what do you want your guest do do? 1126014676 M * VooDooMaster Greek0: I tried --interface vs1=eth0: and another time vs2=eth0: and the second alias does not show 1126014706 M * VooDooMaster Greek0: act as a webserver, the other act as a samba server ... 1126014723 M * Greek0 I suggest just tuning /etc/vservers/*/interfaces once you have the vservers setup 1126014729 M * Greek0 ok, for the webserver it's pretty easy 1126014739 M * Greek0 (I have that setup too) 1126014783 M * VooDooMaster Greek0: tell me ... 1126014906 M * Greek0 http://paste.debian.net/1810 1126014945 M * Greek0 ok. now how it works. normally when you think about networking, you are thinking about interfaces, right? 1126015054 M * VooDooMaster Greek0: Ah ... cool ... thanks ... 1126015073 M * VooDooMaster Greek0: can you emerge stuff in this vserver as usual? 1126015100 M * Greek0 no 1126015160 M * VooDooMaster Greek0: ah .. within my setup this should be possible ... 1126015179 M * Greek0 the guest can only access the dummy ip 10.0.0.2, and all traffic sent there is discarded. The only way to get a connection is to be connected from the outside on port 88 1126015227 M * Greek0 I have a second interface configured, but it's disabled by default 1126015233 J * ntrs ~ntrs@Dardeene-68.188.50.87.charter-stl.com 1126015238 M * Greek0 that one gives me full net access from the guest 1126015242 M * Greek0 mom 1126015260 M * VooDooMaster Greek0: Yes. I see that in your config files. But I want to setup vservers taht behave yust like real servers. 1126015290 Q * renihs Remote host closed the connection 1126015327 M * Greek0 http://paste.debian.net/1811 1126015330 M * Greek0 my real config 1126015366 M * Greek0 I'm on a 192.168.0.0/24 subnet, and giving the vserver a 192.168.0.* ip enables it to transfer data over the real network 1126015397 M * Greek0 the disabled file in that dir is just there so the interface isn't created normally. It's only when I want to upgrade the guest 1126015442 M * Greek0 192.168.0.4 is an IP _not_ used on my network by anything. it's also _not_ the ip of the host itself. it's just unused. 1126015490 M * VooDooMaster Greek0: Ok ... that sounds better. that helps a lot ... BUT ... what if I want to give my vservers the same IP as my host? 1126015546 M * Greek0 use NAT on the host 1126015578 M * Greek0 bind the guests to some ips (not in the local network) on dummy interfaces, and enable NAT on the host 1126015613 M * Greek0 (i.e. set the appropriate iptables rules, enable /proc/...net../ipv4_forward or whatever it's called 1126015616 M * VooDooMaster greek0: ok ... that's at least a workaround 1126015617 M * Greek0 ) 1126015626 M * Greek0 hmm. 1126015633 M * Greek0 perhaps you can also set the ip to that of the host 1126015648 M * Greek0 just try it (I would be interrested too if this works) 1126015789 J * matta ~matta@69.93.28.254 1126015794 M * VooDooMaster greek: it works - I tried that before 1126015841 M * Greek0 ah, ok 1126015849 M * Greek0 where was your problem then actually? 1126015854 M * VooDooMaster Greek0: But only once - then you can't create new aliases ... but the new vservers use the existing IF 1126015875 M * Greek0 why new aliases? 1126015888 M * Greek0 why aliases at all? 1126016000 M * VooDooMaster greek0: I want to test a little traffic monitoring ... for that I need different interfaces for the vservers 1126016092 M * Greek0 hmm. no idea then, sorry 1126016313 M * VooDooMaster Greek0: I'll continue trying ... thanks for your help! 1126016338 M * VooDooMaster Greek0: I'll msg you, if I know how it works :) 1126016349 M * Greek0 :) 1126016844 Q * neofutur Ping timeout: 480 seconds 1126016893 M * ntrs How do you all deal with the annoying ssh brute force attacks on your virtual servers? 1126016932 M * Greek0 not at all, currently 1126016979 M * ntrs hmm, I was wondering if someone has a good idea. 1126016989 M * ntrs I tried to use iptables somehow like this: 1126017017 M * ntrs -A INPUT -p tcp --syn --dport 22 -m recent --set --name sshattack 1126017017 M * ntrs -A INPUT -p tcp --dport 22 --syn -m recent --name sshattack --rcheck --seconds 60 --hitcount 8 -j LOG --log-prefix "SSH_brute_force: " --log-level debug 1126017017 M * ntrs -A INPUT -p tcp --dport 22 --syn -m recent --name sshattack --rcheck --seconds 60 --hitcount 8 --rttl -j DROP 1126017026 M * ntrs but it is not working reliably. 1126017051 M * ntrs For some unknown reason it blocks some virtual server users on the first connection. 1126017132 J * neofutur ~neofutur@neofutur.net 1126017143 N * BobR BobR_oO 1126017153 M * ntrs I also found some scripts that would read the log file for failed logins but it cannot be applied in a virtual server environment because it needs to read all vservers' log files not just the one on the host. 1126017815 J * Ben81 ~Ben81@tipi0e.lri.fr 1126017825 M * Ben81 hello 1126017895 M * Ben81 i use the vserver 2.0 version and i search the file where to add the flag CAP_NET_RAW (like in the old version for 2.4). 1126017975 M * Ben81 anyone ? 1126018159 Q * litage Ping timeout: 480 seconds 1126018205 M * Greek0 hmm 1126018231 M * Greek0 the right place is /etc/vservers/<>/[bc]capabilities 1126018242 M * Ben81 ok i try it 1126018317 M * Greek0 http://www.nongnu.org/util-vserver/doc/conf/configuration.html <-- search for [bc]capabilites there 1126018326 M * Greek0 there are 2 links to files that describe what the caps actually mean 1126018401 M * Ben81 ok :) 1126018412 Q * Hollow Ping timeout: 480 seconds 1126018453 M * Greek0 maybe you don't need _NET_RAW, but instead _RAW_ICMP (or whatever it's called) is enough. 1126018463 M * Greek0 the latter is more specific and hence, more secure 1126018482 M * Ben81 i have to change routing table 1126018553 M * Greek0 oh, ok :) 1126018607 M * Ben81 argh : SIOCADDRT: Operation not permitted 1126018755 M * Greek0 check /proc/virtual//status 1126018828 M * Ben81 cat /usr/local/etc/vservers/test/capabilities CAP_NET_RAW 1126018848 M * Greek0 no, capabilities is outdated 1126018877 M * Ben81 UseCnt: 12 1126018877 M * Ben81 Tasks: 5 1126018877 M * Ben81 Flags: 0000000202000010 1126018877 M * Ben81 BCaps: 00000000344c04ff 1126018877 M * Ben81 CCaps: 0000000000000101 1126018879 M * Ben81 Ticks: 0 1126018884 M * Greek0 Ben81: kernel 2.4 or 2.6? 1126018889 M * Ben81 2.6 1126018895 M * Greek0 ok, capabilities is for 2.4 1126018900 M * Ben81 :( 1126018900 M * Ben81 ok 1126018906 M * Ben81 bcapabilities ? 1126018919 J * litage ~nick@203.201.96.60 1126018921 M * Ben81 i have already try but it was the same 1126018938 M * Greek0 yep, NET_RAW in bcapabilities 1126018956 M * Ben81 ok NET_RAW, i have put CAP_NET_RAW :/ 1126018963 M * Greek0 sure you need NET_RAW and not NET_ADMIN? 1126018970 M * Greek0 for the routing stuff? 1126018986 M * Ben81 mmmm 1126018992 M * Ben81 i don't know 1126019003 M * matta Bertl_zZ: ? 1126019006 M * matta ok 1126019009 M * matta cat is out of the bag 1126019013 M * matta www.openvirtuozzo.org 1126019064 M * Greek0 for modification of routing tables you need NET_ADMIN 1126019069 M * Ben81 yes with NET_ADMIN it works , thaanks Greek0 1126019085 M * Greek0 but you'd better be careful with that.. 1126019097 M * Ben81 yes ;) 1126019103 M * Greek0 http://paste.debian.net/1812 1126019109 M * Greek0 from the kernel source 1126019145 M * Greek0 brb.. dinner 1126020037 Q * monrad Quit: Leaving 1126020252 J * Hollow ~Hollow@home.xnull.de 1126020286 M * Hollow Greek0: i commited the swig interface along with a ruby extconf module to svn 1126020358 M * Greek0 Hollow: mm 1126020384 A * Greek0 tries to install debian on his new sparcstation now :) 1126020774 M * SiD3WiNDR :) 1126020997 M * HostingGeek debian r0x0rz 1126021224 M * Greek0 goil.. "The system is coming up. Please wait." <-- first thing I saw from that lovely being on the floor here 1126021808 J * stefani ~stefani@superquan.apl.washington.edu 1126022021 Q * prae Quit: Execute Order 69 ! 1126022480 Q * kas_3 Ping timeout: 480 seconds 1126022500 Q * Ben81 Quit: Leaving 1126023002 Q * nokoya Ping timeout: 480 seconds 1126023271 J * hillct ~hillct@client200-5.dsl.intrex.net 1126023317 M * hillct morning all 1126023330 M * VooDooMaster ntrs: Didi you try DenyHosts ... it's an nice py-script ... but I think you hav to run it on every vserver OR you start it on the host and modify it to read the logfiles of the vservers 1126023390 M * ntrs VooDooMaster, yes I have seen it but I don't think it will be trivial to modify it to read the logfiles in all the vservers. 1126023456 M * ydupont hello all 1126023459 M * VooDooMaster ntrs: hmmm ... you could start multiple instances - each monitoring one vservers /var/log/messages 1126023494 M * ydupont i've been reading your messages of the day & I have a more or less similar question regarding routing... 1126023498 M * VooDooMaster ntrs: shouldn't consume too much power ... how many vservers do you have on one box? 1126023507 M * ntrs VooDooMaster, :) yes but what do you think the load will be on that host just from the overhead of so many instances alone? 1126023513 M * ntrs at least 30 1126023516 M * ydupont is shere a routing expert here ? 1126023571 M * VooDooMaster ntrs: should work with 30 I think ... just give it a try ... I have it running on a test machine (without vservers) and it's doing a great job. 1126023591 M * VooDooMaster yudupont: just ask your question ... 1126023605 M * VooDooMaster ydupont: just ask your question ... 1126023641 M * hillct was there a change to the behavior of sync-timeout between 1.95 and 2.01? 1126023665 M * hillct after upgrading, I'm seeing timeout warnings where I didn't before 1126023682 M * hillct on vserver shutdown 1126023715 M * ydupont VooDooMaster: is there a way to FORCE the routing of a packet on one interface, even if the @IP to join belongs to another interface on the very same host ? 1126023779 M * ydupont VooDooMaster: the usual IP rule doesn't work here, because the Rule 0 seems to have more priority 1126023829 M * VooDooMaster hillct: me2 - but I think Hollow is working on that 1126023856 M * hillct so this is a known problem? or can it be addressed by modifying the configuration? 1126023907 J * nokoya young@hi-230-82.tm.net.org.my 1126023970 M * hillct VooDooMaster is it a function of the userspace tools or the kernel patches? I'm still learning about the general operation of vservers in this regard 1126024001 M * VooDooMaster ydupont: sounds odd ... maybe someone can help you ... maybe you can do the trick with iptables and snat/dnat rules ... don't know ... sorry 1126024038 M * ydupont VooDooMaster: In case of you're wondering why the hell I need to do that it's because the vservers (the realservers) are acceded by a LVS director. AND my vservers (the realservers) are themselve clients of the LVS ... so the natural routing rules made a direct answer beetween the 2 vservers (they are reachable) , but They don't pass the LVS, and so are not demasqueraded 1126024045 M * VooDooMaster hillct: Hollow only told me that it's a known problem and that he's on to it 1126024100 M * ydupont VooDooMaster: I've been trying all sorts of all solutions all the day... snat/dnat ; fwmark; supress local tables... Absolutely no success :( 1126024101 M * VooDooMaster ydupont: Wow ... nice setup ... 1126024158 M * VooDooMaster ydupont: I have no idea what else to try ... sorry ... 1126024166 M * ydupont VooDooMaster: yes, works very nice, except when the LVS directs the server on the same hosts :( 1126024182 M * ydupont VooDooMaster: thanks anyway, have a nice day 1126024187 M * VooDooMaster ydupont: Maybe one of the other guys can help you? 1126024229 M * hillct VooDooMaster thanks 1126024245 M * VooDooMaster hillct: no problem ... ;) 1126024280 M * ydupont I've searched all the days on mailing lists archives, lvs howto & iproute... it is probably possible to do it, but I still haven't find what I'm looking for (huu ? sounds familiar) 1126024323 M * ydupont VooDooMaster: I think I'll post on the mailing list 1126024358 M * hillct ydupont in my configuration, I use a mysql server run on the host box and accessed by the guest vservers 1126024360 M * VooDooMaster ydupont: good luck! 1126024427 M * ydupont hillct: how they acces each other ? direct routing ? 1126024442 M * hillct ydupont it's my understanding though that access from one vserver to another should be handled using the external IP of each guest, in that such routing will always be treated as a loopback regardless 1126024476 M * hillct This was explained to me by Bertl_zZ who's probably the guy who'd know best 1126024495 M * ydupont hillct: well it's because you have the rule 0 with all the @IP adress 1126024499 M * Hollow VooDooMaster: well, it's known, but i'm not directly working at it 1126024506 M * ydupont hillct: this is the table 0 : local 1126024516 M * Hollow anyway, it works with the timeout... 1126024543 M * ydupont hillct: as this table has a priority 0 it seems to precede all the ip rules I can make... :( 1126024576 M * VooDooMaster Hollow: Last time we talked I think you said somethin like "I know - I'm going to look at it" ... ;) ... but you're right ... it's working ;) 1126024625 M * VooDooMaster Hollow: Can I setup interfaces for a vserver so that inside the vserver ifconfig shows no IF but the vserver has full net access? 1126024648 M * Hollow well... i looked at it quite often.. but util-vserver is a messm and it's not quite trivial to fix.. 1126024668 M * Hollow VooDooMaster: touch /etc/vservers/foo/interfaces/0/nodev 1126024674 M * hillct Hollow my concern is a problem that I experienced under 1.95 and tools 206 where if not peroperly configured, a vserver shutdown durring a host system shutdown would hang the host system 1126024676 M * VooDooMaster hollow: I do fully understand that ... I took al look at it, too ,) 1126024676 M * Hollow ehm 1126024677 M * Hollow wrong 1126024724 M * Hollow just don't create /etc/vservers/foo/interfaces/0/alias iirc 1126024755 M * VooDooMaster hollow: ah ... ok ... 1126024776 M * hillct at the moment I don't believe the use of nodev is well documented on the flower page 1126024799 M * hillct It was pointed out to me and I squireled it away in my notes 1126024807 M * VooDooMaster hollow: is there a good(!) docu on what I can do with the interfaces? I looked at the flower page - but it was not really helpful for my needs 1126024820 M * Hollow i use nodev on my server, works quite well 1126024831 M * hillct as do I 1126024836 M * hillct just not well documented 1126024892 Q * nokoya Ping timeout: 480 seconds 1126024949 M * hillct The flower page is really beginning to annoy me. I was thinking of creating a DB driven interface where each configuration node could be viewed and commented upon and its' children viewed and commented upon, etc etc 1126024992 M * hillct along the same lines, at some point I plan to create a perl object class for vservers, allowing configuration and management 1126024996 M * hillct at some point 1126025037 M * hillct Today though, will be my first attempt at an FC4 guest installation 1126025059 M * hillct so many things to do, and so little time 1126025163 Q * matta Quit: Hey! Where'd my controlling terminal go? 1126025260 J * nokoya young@hi-230-82.tm.net.org.my 1126026031 J * menomc ~amery@200.75.27.24 1126026137 Q * mnemoc Ping timeout: 480 seconds 1126026137 N * menomc mnemoc 1126026312 M * ydupont bye all ! 1126026316 Q * ydupont Quit: Leaving 1126026442 M * hillct where are the vserver build --flags values documented? 1126026569 N * Bertl_zZ Bertl_oO 1126026890 J * matta ~matta@69.93.28.254 1126026912 J * monrad ~monrad@213083190134.sonofon.dk 1126027167 M * daniel_hozac hillct: vserver - build --help 1126027202 M * hillct didn't see the flags enumerated 1126027205 M * hillct will check again 1126028827 J * mef ~mef@targe.CS.Princeton.EDU 1126028996 M * Greek0 Vudumen: ping? 1126029148 M * HostingGeek Greek0: that isn't how you ping 1126029149 M * HostingGeek try 1126029156 M * HostingGeek /ping Vudumen 1126029169 M * HostingGeek or if your crazy: 1126029176 M * HostingGeek /ping #vserver 1126029653 J * nayco ~nayco@lns-th2-8-nan-82-249-47-216.adsl.proxad.net 1126029668 J * Aiken_ ~james@tooax8-100.dialup.optusnet.com.au 1126029679 M * nayco 'llo, all :) !! 1126029716 M * HostingGeek lo 1126029967 Q * logger Ping timeout: 480 seconds 1126029981 M * nayco brb 1126030055 Q * Aiken Ping timeout: 480 seconds 1126030104 Q * nayco Quit: Bonne nuit ! 1126030451 J * prae ~benjamin@sherpadown.net 1126031330 J * logger ~rs@vds.pas-mal.com 1126031782 M * mef is there a way to figure out what the vserver scheduler parameters are for a given context? 1126032799 M * hillct well, it looks like I'm going to have to redo my vserver build process (previously skeleton then yum --installroot=/vservers/blah") to accomodate new dependencies 1126032824 M * hillct a while back someone posted a vserver build process for FC4 guests 1126032835 A * hillct pokes around the Wiki 1126032957 M * TheSeer hillct: apply the build for fc3 and just replace 3 by 4 1126032972 M * TheSeer s/build/howto/ 1126033059 M * hillct I really liked to be able to do it with stock yum, in FC2. Oh well 1126033248 M * TheSeer you got 2.0x and tools? 1126033260 J * lilo_ ~lilo@lilo.usercloak.oftc.net 1126033261 M * TheSeer you have a distro directory which is labeld fc3 1126033284 M * TheSeer you just copy that over to fc4 and adjust the mirror pointers to be fc4 1126033299 J * jpacheco ~justin@CPE0050180cd14b-CM0f0099806976.cpe.net.cable.rogers.com 1126033302 M * TheSeer then specifiy fc4 as your distro of choice when running build 1126033305 Q * lilo Ping timeout: 480 seconds 1126033309 M * jpacheco hey guys 1126033318 M * jpacheco i have a small issue with vserver 1126033335 M * jpacheco whenever i do man in a vserver 1126033342 M * jpacheco it craps out 1126033346 M * jpacheco anyone know why? 1126033350 M * TheSeer define craps out 1126033380 M * jpacheco spits out some stuff and exits 1126033391 M * TheSeer now we know everything :-P 1126033395 M * jpacheco man ls 1126033395 M * jpacheco Error executing formatting or display command. 1126033395 M * jpacheco System command (cd /usr/share/man && (echo ".ll 13.0i"; echo ".nr LL 13.0i"; echo ".pl 1100i"; /bin/gunzip -c '/usr/share/man/man1/ls.1.gz'; echo ".\\\""; echo ".pl \n(nlu+10") | /usr/bin/gtbl | /usr/bin/nroff -Tascii -c -mandoc | /usr/bin/less) exited with status 1. 1126033398 M * jpacheco No manual entry for ls 1126033399 M * jpacheco lol 1126033436 M * TheSeer hmm.. interesting ;> i haven't seen that one before 1126033440 M * jpacheco i think the problem is related to the fact that i can't pass mysql passwords on the command line 1126033451 M * jpacheco same problem i mean 1126033467 M * jpacheco tty or vtty 1126033469 M * jpacheco i dunno 1126033494 M * jpacheco are you able to execute a man command in your vservers? 1126033980 M * hillct ok, so unfortunately apt can't be built for x86_64 1126033989 M * hillct so we use yum 1126034000 M * hillct is there documentation of the build method yum ? 1126034009 M * hillct which i guess uses vyum 1126034027 M * hillct where do the config files go? 1126034040 M * hillct ./distributions/fc4/yum/ what? 1126034424 M * jpacheco anyone running gentoo in a vserver here? 1126035453 M * jpacheco ok 1126035459 M * jpacheco i figured out the problem 1126035471 M * jpacheco if anyone is interested in knowing the answer 1126035486 M * jpacheco "/dev/tty" in my vservers 1126035495 M * jpacheco was not set properly 1126035517 M * jpacheco the major and minor numbers were off 1126035530 M * jpacheco so i did mknod /dev/tty c 5 0 1126035668 M * Hollow jpacheco: which stage did you use? 1126035679 M * jpacheco stage 3 i guess 1126035683 M * jpacheco whatever is up on gentoo 1126035693 M * jpacheco they have a special vserver stage 1126035700 M * Hollow yeah, i created it ;) 1126035706 M * jpacheco ahhh 1126035708 M * jpacheco there we go 1126035718 M * jpacheco oh 1126035720 M * jpacheco wait 1126035731 M * jpacheco u made that package on gentoo? 1126035734 M * Hollow yup 1126035737 M * jpacheco or you created your own /dev/tty 1126035741 M * jpacheco oh wow 1126035741 M * jpacheco lol 1126035755 M * Hollow i'm responsible for all vserver things in gentoo currently 1126035762 M * jpacheco ah, ic 1126035773 M * jpacheco so why did you tty differ from mine? 1126035777 M * Hollow along with DaPhreak 1126035784 M * jpacheco cause i have a debian vserver also 1126035798 M * jpacheco and its /dev/tty was set to 5 0 1126035798 M * Hollow how did it differ? 1126035809 M * Hollow instead of? 1126035811 M * jpacheco in all my gentoo vservers 1126035825 M * jpacheco tty was set to 16 null 1126035843 M * Hollow hmm 1126035861 M * jpacheco as a result 1126035863 M * jpacheco man doesn't work 1126035868 M * jpacheco slappasswd 1126035873 M * jpacheco and mysql -p 1126035891 M * Hollow that's odd 1126035901 M * mnemoc gentoo is odd 1126035904 M * Hollow but i use my stages too, never had that problem before.. 1126035911 M * jpacheco huuum 1126035918 M * jpacheco sounds like its just me then 1126035920 M * jpacheco which is weird 1126035929 M * jpacheco cause i never touch /dev/tty 1126035938 M * Hollow which stage tar exactly did you use? 1126036006 M * jpacheco should be this one 1126036010 M * jpacheco stage3-i686-2005.0.tar.bz2 12-Jun-2005 01:02 82M 1126036041 M * Hollow well... all stage3 are named like this... ;) did you dl it from dev.gentoo.org/~hollow ? 1126036051 M * jpacheco yeah 1126036054 M * Hollow ok 1126036076 M * jpacheco just followed the howto 1126036130 M * Hollow i'm looking at it, though there will be new stages in a few days 1126036148 M * jpacheco for 2005.1? 1126036150 M * Hollow the latest uploaded stage is known to have some minor bugs 1126036166 M * Hollow well, we decided to not follow the main release cycle, but kind of 2005.1 yeah 1126036173 M * jpacheco ahh ic 1126036198 M * jpacheco so what is to be included or fixed in this new release? 1126036208 M * Hollow at least until things have proven stable we'll have shorter release cycles 1126036220 J * dddd44 U2FsdGVkX1@tor-irc.dnsbl.oftc.net 1126036245 M * Hollow well, there are some fixes in baselayout-vserver and the profiles have been reviewed as well, a new minimal profile has been added etc 1126036251 M * Hollow there will be minimal stages too 1126036273 M * jpacheco ic 1126036302 M * Hollow but as i said, many things are more or less cosmetic, the current stage should work quite well 1126036315 M * jpacheco what would be nice, i don't know if this exists yet, but a howto on getting gentoo vservers to be their smallest (in terms of size) 1126036340 M * jpacheco my vservers are like 600MB 1126036358 M * Hollow jpacheco: sounds like you're using an own portage tree for every guest 1126036367 M * jpacheco i did at first 1126036374 M * jpacheco but now i share it 1126036381 M * Hollow good :) 1126036383 M * jpacheco lol 1126036385 M * jpacheco but still 1126036389 M * jpacheco its very big 1126036398 M * jpacheco compaired to debian 1126036432 M * jpacheco and perhaps at setup 1126036441 M * Hollow my smallest gentoo vserver is 257MB, setup with my stage3 + some small apps 1126036452 M * jpacheco that details the steps needed to remove gcc 1126036456 M * jpacheco and all that other stuff 1126036467 M * jpacheco that would be sweeet 1126036477 M * michal there is not vserver for 2.6.13 yet, right ? 1126036488 M * mnemoc 2.0.1-pre2 1126036505 M * Hollow jpacheco: after the new stages have been releases we'll take a look at external packet management for gentoo, so you won't need gcc in your guests probably 1126036530 M * jpacheco nice 1126036565 M * jpacheco cause with a master gentoo vserver doing all the compiling etc, the guests shouldn't need all those dev tools 1126036568 M * Hollow but imo this will be a lot of work 1126036581 M * jpacheco lol, no doubt 1126036599 M * Hollow especially as many ebuilds don't honor $ROOT 1126036640 M * Hollow but we'll see 1126036682 M * Hollow jpacheco: the new stage3 based on the minimal profile has only 44MB as tbz2 1126036692 M * michal i cannot find 2.0.1-pre2 on vserver site... 1126036693 M * jpacheco niiiice 1126036735 M * Hollow jpacheco: http://phreak.xnull.de/gentoo/vserver/stages/ 1126036747 M * Hollow if you're brave ;) 1126036754 M * Hollow they are not tested in any way yet ;) 1126036778 M * jpacheco lol 1126036798 M * jpacheco im still making sure 2.6.13-vs-2.0.1 is stable enough 1126036803 M * jpacheco :) 1126036806 M * Hollow heh 1126036830 M * jpacheco i wanted to try out inotify, but its not showing up in /dev 1126036862 M * jpacheco cd /etc/ssl/ 1126036863 M * jpacheco ls 1126036874 M * jpacheco openssl req -config /etc/ssl/openssl.cnf \ 1126036874 M * jpacheco -new -x509 -nodes -out ldap.pem \ 1126036876 M * jpacheco -keyout /etc/openldap/ssl/ldap.pem -days 999999 1126036882 M * jpacheco oops 1126036885 M * Hollow :) 1126036889 M * jpacheco losin my mind here 1126037033 Q * maharaja Read error: No route to host 1126037045 A * SNy is picking up a rogue mind 1126037047 M * SNy hm 1126037357 M * Hollow jpacheco: btw, i use inotify, and nothing shows up in my /dev 1126037367 M * jpacheco ah 1126037371 M * jpacheco so im not the only one 1126037385 M * jpacheco i can't find anything on google though 1126037421 M * Hollow well... as gamin uses inotify, and i have gamin running with my kde, it seems to work anyway 1126037576 J * kas_3 ~dhb55@tor-irc.dnsbl.oftc.net 1126037744 M * jpacheco huuuuum 1126037747 M * jpacheco interesting 1126037845 Q * dddd44 Ping timeout: 480 seconds 1126038165 M * SNy I have a mind there to give away 1126038177 M * SNy looks like your name was on it, jpacheco 1126038813 M * Hollow SNy: i'll take it... mine has gone away years ago.. ;) 1126039049 Q * weasel Quit: Changing server 1126039057 J * weasel weasel@asteria.debian.or.at 1126039631 J * dddd44 ~dhb55@tor-irc.dnsbl.oftc.net 1126039670 Q * kas_3 Ping timeout: 480 seconds 1126039803 M * jpacheco Hollow: does your vservers have trouble shuting down? 1126039812 P * stefani I'm Parting (the water) 1126039816 M * Hollow jpacheco: except for the timeout, no 1126039843 M * jpacheco yeah timeout is what i get too 1126039846 M * jpacheco why is that? 1126040133 J * kas_3 ~dhb55@tor-irc.dnsbl.oftc.net 1126040449 M * hillct I'm seeing a problem with my new FC4 guest enviroment 1126040459 M * hillct it always hangs starting sendmail 1126040512 M * hillct this appears to be a networking issue of some kind as when I shut down, the vserver I get 1126040513 M * hillct RTNETLINK answers: Cannot assign requested address 1126040520 Q * dddd44 Ping timeout: 480 seconds 1126040645 Q * kas_3 Ping timeout: 480 seconds 1126040645 J * dddd44 ~dhb55@tor-irc.dnsbl.oftc.net 1126040833 M * hillct yeah, the network isn't coming up at all 1126041071 M * hillct I'm sure I've forgotten some small but key step 1126041175 Q * dddd44 Ping timeout: 480 seconds 1126041322 J * dddd44 ~dhb55@tor-irc.dnsbl.oftc.net 1126041419 J * kas_3 ~dhb55@tor-irc.dnsbl.oftc.net 1126041711 Q * jpacheco Quit: BitchX-1.1-final -- just do it. 1126041743 M * hillct my loopback (nodev) device appears to be up, but the external (eth0) device isn't 1126041820 Q * dddd44 Ping timeout: 480 seconds 1126041866 J * dddd44 ~dhb55@tor-irc.dnsbl.oftc.net 1126042080 Q * kas_3 Ping timeout: 480 seconds 1126042191 J * mess-mate ~mess-mate@lns-vlq-7-lil-82-254-206-41.adsl.proxad.net 1126042216 Q * mess-mate Quit: 1126042238 Q * mef Remote host closed the connection 1126043600 M * SNy Hollow: hm, as the original owner doesn't seem to keen on getting it back 1126043618 M * Hollow :P 1126043879 M * Hollow off for now, maybe back later 1126044424 Q * Aiken_ Quit: Leaving 1126044559 J * Aiken ~james@tooax8-100.dialup.optusnet.com.au 1126044771 J * kas_3 U2FsdGVkX1@tor-irc.dnsbl.oftc.net 1126045125 Q * dddd44 Ping timeout: 480 seconds 1126045138 J * dddd44 ~dhb55@tor-irc.dnsbl.oftc.net 1126045234 J * mess-mate ~mess-mate@lns-vlq-7-lil-82-254-206-41.adsl.proxad.net 1126045277 Q * mess-mate Quit: 1126045285 Q * kas_3 Ping timeout: 480 seconds 1126045405 Q * prae Quit: Pwet 1126045720 Q * weasel kinetic.oftc.net arion.oftc.net 1126045720 Q * nox kinetic.oftc.net arion.oftc.net 1126045720 Q * obi kinetic.oftc.net arion.oftc.net 1126045720 Q * ntrs kinetic.oftc.net arion.oftc.net 1126045720 Q * Hunger kinetic.oftc.net arion.oftc.net 1126045720 Q * monrad kinetic.oftc.net arion.oftc.net 1126045720 Q * matta kinetic.oftc.net arion.oftc.net 1126045720 Q * neofutur kinetic.oftc.net arion.oftc.net 1126045720 Q * VooDooMaster kinetic.oftc.net arion.oftc.net 1126045720 Q * click kinetic.oftc.net arion.oftc.net 1126045720 Q * Tenchi-san kinetic.oftc.net arion.oftc.net 1126045720 Q * eugenesan kinetic.oftc.net arion.oftc.net 1126045720 Q * ag-2 kinetic.oftc.net arion.oftc.net 1126045720 Q * TheSeer kinetic.oftc.net arion.oftc.net 1126045720 Q * mugwump kinetic.oftc.net arion.oftc.net 1126045720 Q * pusling kinetic.oftc.net arion.oftc.net 1126045720 Q * Aiken kinetic.oftc.net arion.oftc.net 1126045720 Q * hillct kinetic.oftc.net arion.oftc.net 1126045720 Q * eyck kinetic.oftc.net arion.oftc.net 1126045720 Q * Loki|muh kinetic.oftc.net arion.oftc.net 1126045720 Q * michal kinetic.oftc.net arion.oftc.net 1126045720 Q * toidinamai kinetic.oftc.net arion.oftc.net 1126045720 Q * virtuoso kinetic.oftc.net arion.oftc.net 1126045720 Q * revenger_ kinetic.oftc.net arion.oftc.net 1126045720 Q * duckx kinetic.oftc.net arion.oftc.net 1126045720 Q * dsoul kinetic.oftc.net arion.oftc.net 1126045720 Q * BWare kinetic.oftc.net arion.oftc.net 1126045720 Q * meebey kinetic.oftc.net arion.oftc.net 1126045720 Q * locksy kinetic.oftc.net arion.oftc.net 1126045720 Q * Bertl_oO kinetic.oftc.net arion.oftc.net 1126045720 Q * cereal kinetic.oftc.net arion.oftc.net 1126045720 Q * sladen kinetic.oftc.net arion.oftc.net 1126045720 Q * derbien kinetic.oftc.net arion.oftc.net 1126045720 Q * case kinetic.oftc.net arion.oftc.net 1126045720 Q * Beave kinetic.oftc.net arion.oftc.net 1126045720 Q * janra_ kinetic.oftc.net arion.oftc.net 1126045720 Q * lilo_ kinetic.oftc.net arion.oftc.net 1126045720 Q * logger kinetic.oftc.net arion.oftc.net 1126045720 Q * nokoya kinetic.oftc.net arion.oftc.net 1126045720 Q * Hollow kinetic.oftc.net arion.oftc.net 1126045720 Q * litage kinetic.oftc.net arion.oftc.net 1126045720 Q * Blissex kinetic.oftc.net arion.oftc.net 1126045720 Q * micah_ kinetic.oftc.net arion.oftc.net 1126045720 Q * tchan kinetic.oftc.net arion.oftc.net 1126045720 Q * wibble kinetic.oftc.net arion.oftc.net 1126045720 Q * SiD3WiNDR kinetic.oftc.net arion.oftc.net 1126045720 Q * BobR_oO kinetic.oftc.net arion.oftc.net 1126045720 Q * daniel_hozac kinetic.oftc.net arion.oftc.net 1126045720 Q * [MUPPETS]Gonzo kinetic.oftc.net arion.oftc.net 1126045720 Q * HostingGeek kinetic.oftc.net arion.oftc.net 1126045720 Q * alexx kinetic.oftc.net arion.oftc.net 1126045720 Q * stupidawy kinetic.oftc.net arion.oftc.net 1126045720 Q * DaCa kinetic.oftc.net arion.oftc.net 1126045720 Q * mcp kinetic.oftc.net arion.oftc.net 1126045731 J * kas_3 ~dhb55@ouagadougou.absurdistan.org 1126045760 J * nox ~nox@nox.user.oftc.net 1126045760 J * weasel weasel@weasel.noc.oftc.net 1126045760 J * DaCa ~danny@mail.limehouse.org 1126045760 J * Beave ~beave@vistech.org 1126045760 J * janra_ janra@paradox.homeip.net 1126045760 J * mcp ~hightower@wolk-project.de 1126045760 J * sladen paul@starsky.19inch.net 1126045760 J * case ~case@donpanic.faveve.uni-stuttgart.de 1126045760 J * cereal ~cereal@217.20.124.153 1126045760 J * Bertl_oO ~herbert@212.16.62.52 1126045760 J * locksy ~locksy@mrtg.sisgroup.com.au 1126045760 J * meebey meebey@booster.qnetp.net 1126045760 J * BWare ~bware@office.intouch.net 1126045760 J * dsoul darksoul@vice.ii.uj.edu.pl 1126045760 J * duckx ~Duck@mna75-1-81-57-39-234.fbx.proxad.net 1126045760 J * stupidawy foo@198.77.239.131 1126045760 J * revenger_ ~joe@bulldog.infosys.de 1126045760 J * virtuoso ~s0t0na@80.253.205.251 1126045760 J * alexx ~alexx@proxy.ikse.net 1126045760 J * toidinamai ~frank@toidinamai.de 1126045760 J * obi ~obi@asus.saftware.de 1126045760 J * HostingGeek ~m00@200.48.233.220.exetel.com.au 1126045760 J * [MUPPETS]Gonzo ~gonzo@langweiligneutral.deswahnsinns.de 1126045760 J * daniel_hozac ~daniel@c-6f1472d5.010-230-73746f22.cust.bredbandsbolaget.se 1126045760 J * michal ~michal@michal.usercloak.oftc.net 1126045760 J * BobR_oO ~georg@212.16.62.52 1126045760 J * SiD3WiNDR luser@bastard-operator.from-hell.be 1126045760 J * wibble ~tim@sophie.wobb1e.co.uk 1126045760 J * tchan ~tchan@c-24-13-81-164.hsd1.il.comcast.net 1126045760 J * eyck eyck@81.219.64.71 1126045760 J * micah_ micah@micha.hampshire.edu 1126045760 J * Blissex ~Blissex@82-69-39-138.dsl.in-addr.zen.co.uk 1126045760 J * litage ~nick@203.201.96.60 1126045760 J * Hollow ~Hollow@home.xnull.de 1126045760 J * hillct ~hillct@client200-5.dsl.intrex.net 1126045760 J * nokoya young@hi-230-82.tm.net.org.my 1126045760 J * logger ~rs@vds.pas-mal.com 1126045760 J * lilo_ ~lilo@lilo.usercloak.oftc.net 1126045760 J * Aiken ~james@tooax8-100.dialup.optusnet.com.au 1126045760 J * Loki|muh_ loki@satanix.de 1126045760 J * Aiken_ ~james@tooax8-100.dialup.optusnet.com.au 1126045760 T * kinetic.oftc.net http://linux-vserver.org/ | latest stable 2.0, 2.0.1-pre2, 1.2.10, 1.2.11-rc1, devel 2.1.0-pre5 | Dear Mr Foo, we can not answer your question within 1 minute it is simply impossible... please give us at least 1 hour 1126045768 J * derbien ~derbien@whiterabbit.nbmc.de 1126045768 J * monrad ~monrad@213083190134.sonofon.dk 1126045768 J * matta ~matta@69.93.28.254 1126045768 J * neofutur ~neofutur@neofutur.net 1126045768 J * VooDooMaster VooDoo@topas.informatik.uni-ulm.de 1126045768 J * click click@dsl-static-122-208.aal.tiscali.no 1126045768 J * Tenchi-san ask@208.53.159.170 1126045768 J * eugenesan eugenesan@bzq-219-189-203.cablep.bezeqint.net 1126045768 J * ag-2 ag@muaddib.roxor.cx 1126045768 J * TheSeer ~theseer@border.office.salesemotion.net 1126045768 J * mugwump ~samv@watts.utsl.gen.nz 1126045768 J * pusling pusling@195.215.29.124 1126045771 J * ntrs ~ntrs@Dardeene-68.188.50.87.charter-stl.com 1126045771 J * Hunger Hunger.hu@Hunger.hu 1126046050 Q * Doener Ping timeout: 480 seconds 1126046094 J * Doener ~doener@p54876427.dip.t-dialin.net 1126046155 Q * dddd44 Ping timeout: 480 seconds 1126046162 Q * Aiken Ping timeout: 480 seconds 1126046198 J * dddd44 debian-tor@tor-irc.dnsbl.oftc.net 1126046270 Q * kas_3 Ping timeout: 480 seconds 1126046876 Q * Blissex Remote host closed the connection 1126047704 J * nayco ~nayco@lns-th2-8-nan-82-249-47-216.adsl.proxad.net 1126047812 Q * weasel Write error: connection closed 1126048213 N * Bertl_oO Bertl 1126048217 M * Bertl evening folks! 1126048261 M * nayco helo ! 1126048277 M * nayco Bertl: Mail :) !!! 1126048319 M * nayco well, mailS... 1126048321 M * Bertl hey nayco! 1126048340 M * Bertl xfs quota? 1126048402 M * hillct evening Bertl 1126048417 M * Bertl evening hillct! everything fine? 1126048432 M * nayco yep, and netatalk, and one about a few bugs. These last two on the list. 1126048434 M * nayco ;-) 1126048440 M * hillct well, I worked out the problem with the util-vserver RPMs 1126048462 M * hillct fought my way through installing an FC4 guest using yum 1126048473 M * hillct but I'm having a routing problem 1126048498 M * hillct guest eth0 is not routing outside of the host box 1126048538 M * hillct guests created with previous tools didn't require any special steps to get external routing 1126048549 M * hillct I'm sure it's a simple step I'm missing 1126048690 M * Bertl nayco: okay, you have a setup for xfs quota testing? 1126048720 M * nayco well, I'll have one tomorrow ;-) 1126048723 M * Bertl hillct: I guess you simply have a different setup now ... 1126048744 M * Bertl hillct: could you upload a short description of your network setup somewhere` 1126048750 M * Bertl s/`/? 1126048769 M * hillct sure 1126048771 M * nayco Bertl: One question: Why XFS only ? Disk quota are implemented fs-dependent in the kernel ? 1126048822 M * Bertl unfortunatley yes, especially for non linux fielsystems 1126048833 M * Bertl *filesystems 1126048838 Q * monrad Quit: Leaving 1126048842 M * nayco Oh that's sad. 1126048909 M * Bertl yes, especially as there is a pretty robust quota system which is used for ext2/3 and IIRC jfs too 1126048930 M * nayco ok, anyway, for XFS (Which is the best multi-purpose filesystem for servers, I heard), no problem : I can use one of my hosts to test... 1126048975 M * nayco Bertl: Errr... I need a dedicated host ? Oh, yes, I gonna try kernel patches, huh ? 1126049089 M * Bertl yes, we will first narrow down _why_ it doesn't work, then try some patches ... 1126049127 M * nayco Ok. You said ext* have good quota system: XFS's ids bad ? 1126049132 M * nayco -d 1126049201 M * Bertl well don't know much about XFS, it's very complicated and not well integrated into the kernel 1126049209 M * nayco arf... 1126049221 M * Bertl (read: it's a horror to read XFS code, but I will try :) 1126049285 M * nayco ;-) !!! Oh, a co-worker told me today that I could try to set quotas from the host, and that would work inside host using the given filesystem, is it true ? 1126049310 M * nayco ...Should use unified user IDs... no . 1126049312 M * nayco ? 1126049423 M * Bertl yes, it would work 1126049438 M * hillct Bertl http://42.vg/77431 1126049445 M * Bertl nayco: but it will not allow you to modify or query them from the guest 1126049528 M * Bertl hillct: tx 1126049547 M * nayco Well I think is boring : Having to share the same user IDs across the host and vservers is a constraint, and worse if using Winbind to be a member file server for NT, it get impossible. 1126049553 M * hillct brb 1126049596 M * nayco Oh, I forgot "it" and "s" ;-) 1126049665 M * Bertl nayco: yes, as I said, I think we might get it working (at least for separate partitions) 1126049730 M * nayco Well, that would be great to be able to set quotas in/from the vserver, but thinking of it, I don't see how this can work without actually seeing the mountpoint inside the vserver (and not only /dev/hdvx) 1126049795 M * Bertl hillct: okay, can you provide an 'ip addr ls' and 'ip route ls' output of your host system? (you might replace sensitive information by placeholders, e.g. xxx, yyy ...) 1126049797 M * nayco So, I guess setting quotas inside a vserver will look different from setting them on a normal server. Well, I speak about the command line that does it ;-) 1126049828 M * Bertl nayco: no, the idea is always to make it equivalent to the host/real system 1126049882 M * Bertl nayco: first thing you have to show/explain to me is how the xfs quota works from the user perspective ... 1126049887 M * nayco Good point. So, we need a mean to make for example the /home mountpoint visible from the guest, at least virtually ? 1126049948 M * Bertl that is, what the vroot device is for 1126050056 M * nayco er... well... It looks like adding "quota" in the mount options in fstab, then reboot the system (Yes, it's strange, there must be another way), and finally "setquota 102400 102400 0 0 1126050067 M * nayco Is it what you wanted to know ? 1126050128 M * nayco oh, It got cut: "setquota username 102400 102400 0 0 /home" gives username a 100Mb quota on /home 1126050159 M * hillct http://42.vg/77432 1126050164 M * Bertl nayco: what's the 102400 102400 0 0 mean? 1126050172 M * hillct Bertl ^^ 1126050186 M * nayco after that, "repquota /home" list all the quotas for users on /home, tells who is over, whoo is under, etc... 1126050237 M * Bertl hillct: and which one doesn't get 'routed' to the internet? 1126050249 M * hillct .132 1126050294 M * nayco ( http://unixhelp.ed.ac.uk/CGI/man-cgi?setquota+8 ) : setquota username block-softlimit block-hardlimit inode-softlimit inode-hardlimit filesystem 1126050316 M * Bertl hillct: and you're 100% sure that 66.193.14.1 would route that one, and your network is /24 not maybe /25 (which would make it end at 127 :) ? 1126050327 M * nayco Well, that is the same way for all filesystems that support quotas. At least from userland 1126050361 M * Bertl nayco: ah, okay, so it uses similar 'input' as ext2/3 quota 1126050367 M * nayco yep ! 1126050370 M * Bertl good! 1126050371 M * hillct yes. absolutely sure 1126050395 M * nayco That's why I thought it could be the same underlying code kernel-wise ;-) 1126050416 M * Bertl hillct: just to eliminate this possibility, could you assing something below 127, maybe 122 or so? 1126050423 M * Bertl *assign 1126050432 M * nayco assing :D 1126050450 M * hillct not really 1126050463 M * hillct in that I don't have control of the interveening addresses 1126050467 M * Bertl okay, then let's try the following: 1126050472 M * hillct er 1126050479 M * hillct I can swap 123 1126050482 M * hillct I think 1126050494 M * Bertl ping -c 1 -I 66.193.14.121 www.google.com 1126050498 M * hillct that's not a heavily used vserver 1126050498 M * Bertl (on the host) 1126050507 M * Bertl ping -c 1 -I 66.193.14.132 www.google.com 1126050514 M * Bertl (again on the host) 1126050578 M * Bertl nayco: okay, so let's add this "quota" option on the host mount, and do the reboot/remount 1126050594 M * Bertl nayco: best make it a separate partition for testing 1126050607 M * Bertl (is that possible?) 1126050625 M * hillct [root@vmaster vservers]# ping -c 1 -I 66.193.14.121 www.google.com 1126050626 M * hillct bind: Cannot assign requested address 1126050626 M * hillct [root@vmaster vservers]# ping -c 1 -I 66.193.14.132 www.google.com 1126050626 M * hillct bind: Cannot assign requested address 1126050626 M * hillct [root@vmaster vservers]# ping -c 1 -I 66.193.14.132 www.google.com 1126050627 M * hillct bind: Cannot assign requested address 1126050627 M * hillct [root@vmaster vservers]# 1126050639 M * nayco Bertl: well, I gonna fire up a test host tomorrow. 1126050661 M * Bertl hillct: how are you logged on? 1126050668 M * nayco Oh, I can even try on one of my server, it goeas to production only in a couple of days. 1126050709 M * Bertl hillct: via ssh? and do you use the v_ssh wrapper? 1126050710 M * hillct to the host? 1126050746 M * Bertl yes 1126050750 M * hillct ssh 1126050800 M * nayco Bertl: So, you want me to set the quota option on the host, do all that's needed, give a user a quota from the host, mount this partition for use inside a vserver, and then check if the quota is respected inside it ? 1126050816 M * Bertl hillct: what does 'cat /proc/self/ninfo' report? 1126050826 M * hillct on host? 1126050840 M * Bertl hillct: yes, nayco: not exactly ... 1126050855 M * Bertl nayco: just make sure it is turned on (via mount option) 1126050873 M * nayco Ok, only turned on, and nothing else. 1126050876 M * Bertl the next step is to see _what_ fails if we want to use it inside a guest ... 1126050885 M * nayco yep... 1126050891 M * Bertl (so we need a guest for that too) 1126050903 M * hillct info doesn't exist in /proc/self 1126050916 M * Bertl 'ninfo' 1126050917 M * hillct what specific information are we looking for? 1126050928 M * Bertl if that doesn't exist, it's fine ... 1126050939 M * hillct [root@vmaster self]# cat ninfo 1126050940 M * hillct NID: 0 1126050940 M * hillct [root@vmaster self]# 1126050958 M * Bertl okay, so why is ping not able to bind your source ips? 1126050962 M * nayco Bertl: ok. I'll do it tomorrow. Do you have a way to be contacted that could allow use to share tests results fastly ? 1126050979 M * hillct no idea 1126050986 M * hillct er 1126050988 M * hillct wait 1126050992 M * hillct firewall 1126051000 M * hillct d'oh 1126051017 M * nayco ...taht could allow US... 1126051018 M * Bertl nayco: well, let's try to find some time when I'm around tomorrow (around 15-1600 CET) 1126051038 M * nayco er... CET ? Which time is it ? 1126051039 M * hillct has holes for port 22, 53, and 80 1126051055 M * Bertl (it's currently 2 am CET now) 1126051062 M * nayco ok, like here. 1126051088 M * Bertl hillct: could you disable the firewall just for a short moment? 1126051090 M * nayco Well, I do not have IRC at work, do you have an IM (I can use yahoo) 1126051105 M * nayco Or will mail be sufficient ? 1126051116 M * Bertl I have Gnome ICU here, send me a mail with your contact details ... 1126051139 M * nayco what is ICU ? Instant messenging ? 1126051150 M * Bertl something like ICQ :) 1126051154 M * nayco ;-)