1142380855 M * daniel_hozac the other two (RLIMIT_CPU and RLIMIT_SIGPENDING) aren't referenced anywhere else in vserver. 1142380880 M * Bertl yep, that sounds fine 1142381054 J * |coocoon| ~coocoon@p54A06472.dip.t-dialin.net 1142381280 Q * coocoon Ping timeout: 480 seconds 1142381559 Q * pzYsTorM Ping timeout: 480 seconds 1142382461 Q * |coocoon| Quit: KVIrc 3.2.0 'Realia' 1142384993 M * Aiken Bertl oprofile is not looking promising http://pastebin.com/602769 1142385031 M * Bertl hehe .. after opcontrol --start ? 1142385043 M * Aiken yes 1142385064 M * Bertl guess folks @ lkml (especially the oprof ones) would like to get a copy ... 1142385069 M * Aiken I am wonder what 'Instruction fault 4' means 1142385409 J * matta ~matta@c-68-81-35-243.hsd1.pa.comcast.net 1142385535 M * Aiken found a hint indicating it could be a milo problem 1142385552 M * Bertl ah? interesting ... 1142385607 M * Aiken this talks about the function wrperfmon and milo http://lists.debian.org/debian-alpha/1999/03/msg00295.html 1142385615 M * Aiken which seems to be where mine is blowing up 1142385628 M * Aiken I have another milo I can try 1142385669 M * Aiken pc = include/asm/system.h:258 -> __CALL_PAL_RW2(wrperfmon, unsigned long, unsigned long, unsigned long); 1142386015 T * Bertl http://linux-vserver.org/ | latest stable 2.01, 1.2.10, 1.2.11-rc1, devel 2.1.0, exp 2.{0.2,1.1}-rc13 | util-vserver-0.30.210 | libvserver-1.0.2 & vserver-utils-1.0.3 | He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the wiki, and we'll forget about the minute ;) 1142386219 M * daniel_hozac changes? 1142386245 M * daniel_hozac 2.6.16 related only? 1142386310 M * Bertl the s390 stuff I forgot and the sendfile changes 1142386323 M * daniel_hozac hmm. 1142386338 M * Bertl plm is back so I need something to submit for testing :) 1142386363 M * daniel_hozac so no new supported limits or fixed return values? :) 1142386383 M * Bertl hmm, I knew I forgot something ... 1142386408 M * Bertl I'll add the return values, but do you have a patch for the limits yet= 1142386412 M * Bertl s/=/? 1142386548 M * daniel_hozac http://daniel.hozac.com/vserver/delta-valid_limit-fix01.diff 1142386592 M * Bertl only applies for 2.1.1 I presume 1142386663 M * daniel_hozac should apply to 2.0.2 as well. 1142386675 M * Bertl hmm, do we have those limits there? 1142386713 M * daniel_hozac linux-2.6.15.6-vs2.0.2-rc13$ grep -c vx_ipcmsg ipc/mqueue.c 1142386713 M * daniel_hozac 4 1142386713 M * daniel_hozac linux-2.6.15.6-vs2.0.2-rc13$ grep -c vx_locks_ fs/locks.c 1142386713 M * daniel_hozac 8 1142386763 M * Bertl hmm, let's see how it goes in devel first, next round we add it 1142386807 M * daniel_hozac sure. 1142386844 M * Bertl okay, updated the patches in place with the limit fixes 1142386888 M * Bertl i.e. the return value change is in all 4, the limit unlock only in devel 1142387279 M * Bertl hmm .. probably plm is still broken *sigh* 1142387580 M * daniel_hozac looks like. 1142387762 M * Bertl gee, akpm is really damn fast ... 1142387916 M * Bertl Aiken: your alpha is UP, right? 1142387945 M * Aiken not sure, about to go and have a look 1142387955 M * Aiken a reboot seems to have failed 1142387965 M * Bertl UP as not SMP :) 1142387972 M * Aiken yes UP 1142387998 Q * gerrit Read error: Operation timed out 1142387999 M * Bertl okay, great, means we are not seeing something SMP related ... 1142388228 M * Aiken after more reading it still looks like a milo problem, a small number of PAL functions that are not implemented in but are in srm 1142388251 M * Aiken one of those unimplemented function calls is used by oprofile 1142388255 M * Bertl hmm, any work around in sight? 1142388272 M * Aiken everything I found dated from around 1999 1142388289 M * Aiken found a xlt milo from 2000 and same problem with it 1142388393 M * Aiken I know there are some 2.2 kernel based milos but none of them will boot for me 1142388411 M * Aiken the only milo versions I have been able to use are those based on 2.0.X kernels 1142388447 M * Bertl okay, forget about oprofile then, not worth spending more time on it 1142388570 J * Aiken_ ~james@tooax8-191.dialup.optusnet.com.au 1142388858 M * Bertl Aiken_: did you got my last message regarding oprofile? 1142388860 M * Aiken_ chnaging the define in drivers/oprofile/oprof.c that sets timer based profiling instead of register based works :) 1142388864 M * Aiken_ no 1142388872 M * Bertl ah, okay, then nevermind :) 1142388898 M * Bertl Bertl> okay, forget about oprofile then, not worth spending more time on it 1142388899 M * Aiken_ seems going through source code does work sometimes :) 1142388914 Q * Aiken Ping timeout: 480 seconds 1142388916 M * Bertl but if it works now, great! 1142389309 M * Aiken_ I surpose the next step is to run test.sh again with oprofile running 1142389316 M * Aiken_ you want it run as host, guest or both? 1142389369 M * Bertl yes, the oprofile is best used in differential mode here, I'll get you the commands (i.e. I upload a few scripts) 1142389676 M * Bertl http://vserver.13thfloor.at/Stuff/OPROF/prof.sh this is the setup, you want to adjust the kernel path 1142389692 M * Bertl http://vserver.13thfloor.at/Stuff/OPROF/test_prof0.sh and http://vserver.13thfloor.at/Stuff/OPROF/test_prof2.sh 1142389704 M * Bertl will do profiling in xid=0 and xid=2 1142389718 M * Bertl after a profile run, save the profile with: 1142389732 M * Bertl opcontrol --save= 1142389790 M * Bertl we are then interested in something like this: 1142389791 M * Bertl opreport -lt 1 { session:xid0 } { session:xid2 } 1142390815 M * Aiken_ what is --event=CPU_CLK_UNHALTED:50000:0:1:0 1142390853 M * Bertl the recorded event, check with opcontrol --list-events 1142390865 M * Bertl maybe your oprof has different events 1142390878 M * Aiken_ (root@pebbles) opcontrol --list-events 1142390878 M * Aiken_ Using timer interrupt. 1142390905 M * Bertl hmm, maybe remove the --event then 1142393335 J * matt1 ~matta@c-68-81-35-243.hsd1.pa.comcast.net 1142393460 Q * matta Read error: Operation timed out 1142393773 Q * entroposcope Quit: Ping timeout 1142393820 Q * matt1 Ping timeout: 480 seconds 1142394573 M * Aiken_ the final result is only 9,106 lines long opreport -lt 1 { session:host } { session:guest } > /tmp/delme 1142394580 M * Aiken_ time are all wrong 1142394595 M * Aiken_ the host run went a lot longer and the guest run was quicker 1142394613 M * Bertl huh? 1142394662 M * Aiken_ the original compile part of the host was 6:30, with the profiling it was > 15 minutes 1142394679 M * Bertl okay 1142394685 M * Aiken_ the original compile part of the guest was 9 min, with the profiling it was just over 7 minutes 1142394712 M * Bertl cool 1142394733 M * Aiken_ does not make sense 1142394754 M * Bertl not really 1142394856 J * matta ~matta@c-68-32-202-140.hsd1.pa.comcast.net 1142395110 Q * shedi Read error: Connection reset by peer 1142395223 J * entroposcope ~entroposc@user-0c992og.cable.mindspring.com 1142395519 J * Smutje_ ~Smutje@xdsl-87-78-57-234.netcologne.de 1142395644 Q * Smutje Ping timeout: 480 seconds 1142395644 N * Smutje_ Smutje 1142395699 Q * matta Ping timeout: 480 seconds 1142395868 J * gerrit ~gerrit@c-67-160-146-170.hsd1.or.comcast.net 1142395880 J * shedi ~siggi@inferno.lhi.is 1142395938 J * DataCompBoy ~datacompb@217.8.236.1 1142395957 M * DataCompBoy ooo! peoples! I have found what I have searched yesterday: 1142395964 M * DataCompBoy how to log into vservers with only one ssh! 1142395971 M * DataCompBoy http://swapoff.org/LinuxVServer 1142396059 M * Bertl congrats! 1142396185 M * DataCompBoy now i'm close to be happy:) 1142396201 J * Hmmmm ~Hmmmm@221.135.51.19 1142396214 M * DataCompBoy Welcome, Hmmmm! 1142396251 M * Hmmmm hey there 1142396336 M * DataCompBoy How to add such useful links to wiki? doesn't see where to login "( 1142396373 M * Bertl not necessary, just add it, but make use of the preview option (and if you want to be recognized, set a name in the defaults) 1142396454 M * DataCompBoy also, where add it? As I see -- it not well structurized... 1142396478 M * Bertl feel free to improve the structuring .. but think twice :) 1142396584 M * Bertl I would add it to the 'Documentation' page under wiki or so 1142396590 M * Bertl http://linux-vserver.org/Documentation 1142396606 J * matta ~matta@c-68-32-202-140.hsd1.pa.comcast.net 1142396614 M * Bertl or maybe howtos 1142396711 M * DataCompBoy ok, in documentation 1142396741 M * DataCompBoy or in recipes? 1142396842 M * Bertl check, maybe it is already there? 1142396914 M * DataCompBoy no 1142396921 M * Bertl okay, good 1142396923 M * DataCompBoy no links to vcopy, not to vslogin 1142396972 M * Bertl okay, maybe add it twice, once under wiki, the other time under howtos? 1142396994 M * Bertl but personally howtos look good for me 1142397395 M * DataCompBoy 1. http://linux-vserver.org/HowtoCloneVserver 1142397509 Q * matta Ping timeout: 480 seconds 1142397753 M * Bertl DataCompBoy: side note/question: why do you use 0.30.209? 1142397812 M * DataCompBoy shipped with debian etch :) 1142397826 M * Bertl ah, now etch is outdated too? 1142397847 M * DataCompBoy etch is testing 1142397850 M * Bertl I thought that was sarge and sid :) 1142397854 M * DataCompBoy up-to-date is "sid" 1142397864 M * DataCompBoy sarge is stable 1142397870 M * Bertl so sid has 0.30.210? 1142397870 M * DataCompBoy sid is unstable 1142397917 M * DataCompBoy hmm... not, still 209-2 1142397924 M * Bertl fine, so which debian distro has 0.30.210 then? 1142397957 M * DataCompBoy looks like none 1142397973 M * DataCompBoy when 210 awas released? 1142397979 M * Bertl okay, so all are outdated :) 22.jan 1142398145 M * DataCompBoy what's new in 210? 1142398349 M * Bertl you have to check the changelog in the source 1142398361 M * DataCompBoy ok, I have mailed to maintainers 1142398371 M * Bertl but IIRC, support for recent kernel APIs and such 1142398400 M * DataCompBoy what is "IIRC" ? 1142398411 M * Bertl If I Remember Correctly :) 1142398425 M * DataCompBoy :) ok 1142398441 M * DataCompBoy "afair", in other words:] 1142398466 M * DataCompBoy 2. http://linux-vserver.org/HowtoSSHLogin 1142398489 M * DataCompBoy btw: where are checks for spelling in wiki edit? 1142398495 M * DataCompBoy my english toooooo bad 1142398561 M * Bertl hehe, we might add that at some point :) 1142399633 M * Aiken_ http://pastebin.com/603009 1142399641 M * Aiken_ you still want the profile results? 1142399763 M * Bertl yep, maybe we get some info nevertheless 1142399995 M * Bertl Aiken_: you can send it per email as .bz2 if you like 1142400021 M * Aiken_ forget that run, the sample director is empty 1142400105 M * Aiken_ what version oprofile are you using? 1142400131 M * Bertl oprofile-0.9.1-0.1.20060mdk 1142400137 M * Aiken_ grumble, I forgot the save stop that time :( 1142400152 M * eyck morning 1142400173 M * Bertl morning eyck! 1142400224 M * DataCompBoy good, eyck 1142400343 M * DataCompBoy http://linux-vserver.org/HowtoPrivateNetworking 1142400423 M * Bertl hmm, what do you need the ip_forward for? 1142400536 J * Data ~datacompb@217.8.236.1 1142400552 N * Data DataCompBoy2 1142400559 M * DataCompBoy2 pffr, got disconnect 1142400584 M * DataCompBoy2 forwarding from dummy0 looks like also need to have ip_forward enabled 1142400596 M * Bertl did you try that? 1142400625 M * DataCompBoy2 moment, trying 1142400649 M * DataCompBoy2 hmm... looks like you right :) at FORWARD chain nothing catched 1142400654 M * DataCompBoy2 ok, removeing 1142400655 M * DataCompBoy2 :] 1142400661 M * Hollow morning 1142400669 M * Bertl morning Hollow! 1142400678 M * DataCompBoy2 Good, Hollow 1142400825 Q * DataCompBoy Ping timeout: 480 seconds 1142400876 J * ebiederm ~eric@ebiederm.dsl.xmission.com 1142400882 N * DataCompBoy2 DataCompBoy 1142400895 M * Bertl welcome ebiederm! 1142400910 M * ebiederm Hey. 1142400923 M * DataCompBoy YAW, ebiederm! 1142400943 M * ebiederm I was wanting to answer you question about entering a guest a little more. 1142400984 M * Bertl okay, sounds good :) 1142401097 M * ebiederm So after thinking about it I came up with an comletely different way of implementing enter, that handles the security issues beatifully. 1142401145 M * Bertl sounds even better ... 1142401188 M * ebiederm If you can do a ptrace attach across pid spaces you can pick on a process and command do anything you want. 1142401223 M * DataCompBoy Bertl: I have got response from maintainer of util-vserver: 1142401231 M * DataCompBoy "Work is in progress. Most fixes is already there. Actually the Debian version is more updated than the current 0.30.210 version." 1142401253 M * Hollow DataCompBoy: which fixes? 1142401261 M * Bertl DataCompBoy: which maintainer? 1142401269 M * Hollow heh 1142401274 M * eyck DataCompBoy: which debian? 1142401278 M * Hollow lol! 1142401281 M * micah not from me! 1142401287 M * Bertl eyck: broken! 1142401298 M * DataCompBoy Bertl: from maintainer of debian package 1142401300 M * Bertl hey micah! LTNS! 1142401304 M * micah yeah I know! 1142401306 M * DataCompBoy eyck: etch / sid 1142401315 M * micah my work has crushed me 1142401327 M * Bertl micah: seems ola is telling stories again ... 1142401335 M * micah yeah :P 1142401339 M * eyck Ola Lundqvist? 1142401344 M * micah Bertl: well, he is actually sort of right! 1142401345 M * DataCompBoy yup :] 1142401356 M * micah Bertl: because the debian 0.30.209 contains the full shiny8 diff :) 1142401370 M * Bertl micah: yeah, but what about the new kernel ABI? 1142401377 M * micah Bertl: and enrico's 0.30.210 doesn't contain the full shiny8, but of course 0.30.209 has other pieces that are behind 1142401384 M * Hollow iirc shiny7->8 was just mips support, no? 1142401399 M * micah Bertl: exactly, so he just doesn't know the full story it appears 1142401410 M * micah Hollow: and hppa I think 1142401411 M * Bertl Hollow: yeah, so it now works on mips if you replace everything else :) 1142401431 M * Hollow right.. ;) 1142401449 M * micah Bertl: I was hoping Ola would work some on getting 0.30.210 out while I've been so busy with work, but it looks like I will be doing that in the next few days 1142401468 M * micah plus a lot of "fixes" to many things that shouldn't be in there anymore 1142401476 M * Bertl ah, well, nothing new here, let's move on ... 1142401481 M * Bertl micah: TIA 1142401560 M * micah Bertl: yeah! i'm excited to get it out 1142401582 M * micah I wish I could've been around more lately, I wonder what I've been missing 1142401602 M * Bertl tons of good stuff :) 1142401640 M * Bertl ebiederm: so what do you want to do with the ptrace across pid spaces? 1142401713 M * Bertl ebiederm: I'd consider that more a security leak than a feature, but the ovz folks will celebrate :) 1142401716 M * Aiken_ just sending the latest profile results 1142401724 M * Bertl excellent, tx 1142401728 M * ebiederm Bertl: So ptrace if I have the appropriate priveleges let's me turn a process into a puppet. 1142401730 M * Aiken_ am not getting consistant results run to run 1142401742 M * Aiken_ this time the guest is slower than the host 1142401789 M * Aiken_ swap has been turned off for the last couple of runs 1142401800 M * Bertl okay, can't hurt 1142401815 M * Aiken_ takes it out of the equation 1142401821 M * ebiederm Bertl: So ptrace can allow me to fork/exec a shell in a different pid space. 1142401838 M * eyck ebiederm: are you sure you will be able to 'enter' a vserver space that isn't running? 1142401852 M * ebiederm I need to look at an existing implementation of enter to make certain I can get all of the appropriate features. 1142401912 M * ebiederm eyck: Is that a common case? Entering something that isn't running? Actually that should be easier just do all of the work to create the vserver. 1142401931 M * Bertl ebiederm: yes that is a _very_ common case, actually 1142401941 M * Bertl ebiederm: especially if the guest won't start :) 1142401969 M * ebiederm Ah. The classic init=/bin/bash case. 1142401985 M * Bertl yep 1142402020 J * matta ~matta@c-68-32-202-140.hsd1.pa.comcast.net 1142402047 M * micah Bertl: Kir contacted myself and Ola about trying to get debian packages of openvz, interesting 1142402071 M * Bertl well, not quite unexpected ... 1142402149 M * Bertl ebiederm: why are you trying so hard to avoid entering a space explicitely? 1142402153 M * eyck I can't believe it.... Bertl supported my point of view... 1142402180 M * eyck there is only one explanation - he's been replaced by space aliens 1142402222 M * ebiederm Bertl: Actually in this instance, if I can do an operation with ptrace that gives me an existing security model. I don't have a problem with a more convinient kernel interface. 1142402268 M * ebiederm All that happened is that it occured to me this was functional if ptrace worked. 1142402331 M * ebiederm In the model where everything is a done on a per process basis ptrace is not unreasonable, with the appopriate permissions. 1142402433 M * ebiederm So I am primarily after a well defined security model. 1142402656 M * ebiederm One of my concerns with entering a space explicitly is that through mechanism like proc and ptrace by examining a process that has entered it is possible for kernel data structures from the outside to leak in making it easier to escape. 1142402739 M * Bertl that is a reasonable concern, but there are two option here (IMHO) 1142402746 M * ebiederm With ptrace you have to build what you need from what is already there. 1142402750 M * ebiederm Yes? 1142402774 M * Bertl a) make damn sure that the 'enter' process is secure (in userspace or kernel space) 1142402806 M * Bertl b) protect the entering process against 'native' processes 1142402843 M * Bertl a) is what most folks would prefer, b) (at least on 'request') is what I would prefer 1142402864 Q * matta Ping timeout: 480 seconds 1142402893 M * Bertl I'd really love to have three states actually 1142402907 M * Bertl 1) outside (i.e. inside parent pid-space) 1142402945 M * Bertl 2) half way inside the child space (i.e. can act and see, but not be seen or accessed) 1142402966 M * Bertl 3) fully inside the child space (i.e. cannot escape/access outside) 1142403103 M * ebiederm So that is largely what exists with ptrace today. 1142403116 M * Bertl note: this only applies to the pid space, not to the other spaces 1142403128 M * ebiederm The only difference being in the half way inside case there is a poppet process on the inside. 1142403147 M * ebiederm s/poppet/puppet/ 1142403204 M * Bertl just because it was referenced here today ... 1142403206 M * Bertl http://swapoff.org/LinuxVServer 1142403407 M * ebiederm Looks like this is an strong users of enter :) 1142403470 M * Bertl just to give you an idea that this is not only the 'emergency' case 1142403490 M * ebiederm Sure. 1142403493 M * Bertl although I have to admit, that this is not the typical case either 1142403504 M * ebiederm I haven't made any firm decisions yet. 1142403506 M * DataCompBoy have someone runned exim on host to serve all vservers at once? 1142403646 M * ebiederm The full story is while doing some general kernel cleanups another way to do multiple pids per process occurred to me that is possibly cleaner than the wid thing I did. 1142403705 M * ebiederm I'm not quite ready to pursue it but if Kirill can stand to have pids across pids spaces only on demand then the major dead lock on the design discussion is broken. 1142403705 M * Bertl hmm, may I ask again why we need multiple pids at all? 1142403736 M * ebiederm Bertl: So the existing syscall interface works, across pid spaces. 1142403743 M * ebiederm waitpid and the like. 1142403761 M * Bertl do we need that? 1142403775 M * Bertl I mean, why do we need to use existing interfaces for that? 1142403821 J * matta ~matta@c-68-32-202-140.hsd1.pa.comcast.net 1142403830 M * DataCompBoy YAW, matta! 1142403850 M * ebiederm My preference is to use existing interfaces that have already stood the test of time and been fleshed out to where people can do useful work with them. 1142403875 M * ebiederm If we invent our own interfaces we are in for a bumpier ride. 1142403892 M * Bertl I mean it doesn't sound too good to me to say: hey we need to have multiple pids, I know it's complicated but we definitely need it so that we can hack it and use waitpid instead of waitspacepid .... 1142403944 M * ebiederm So that is the interesting thing. I don't think it is complicated. 1142403946 M * Bertl we got the wait and reaper stuff wrong for several month, nobody even noticed ... 1142403990 M * Bertl so IMHO, cross context wait or signalling is nice, but it won't work properly via the existing interfaces 1142404014 M * Bertl why? just because you have to bend backwards to map it 1142404057 M * Bertl in the early stages we came to the conclusion that space:pid would be unique, right? 1142404086 M * ebiederm If I understand the question yes. 1142404103 M * Bertl okay, so how many interfaces user<->kernel do use pids atm? 1142404120 M * ebiederm Bertl: So one of the things that comes with waitpid working is that oddball things like getrusage also work. 1142404207 M * ebiederm capabilites, wait, ioprio, signals, process priorities, a fairly small group. 1142404230 M * Bertl and most of them are not even relevant for cross space stuff 1142404284 M * ebiederm That is a hard call. 1142404321 M * Bertl well, let's assume that 'enter' and 'half-enter' works really nice 1142404323 M * ebiederm They are almost all rarely used management functions. 1142404366 M * Bertl which pid related functions would be required to handle cross space stuff? 1142404396 M * DataCompBoy Bertl: http://vserver.meulie.net/ is dead? :) 1142404430 M * ebiederm Well waitpid so I can later do getrusage(RUSAGE_CHILDREN) on the pid space. 1142404443 M * Bertl DataCompBoy: seems so, but maybe just temporary 1142404477 M * Bertl ebiederm: for what reason? 1142404494 M * Bertl s/reason/purpose/ 1142404530 M * ebiederm HPC applications are generally started by batch schedulers. 1142404536 M * Bertl I mean, what real world scenario would use that? 1142404560 M * ebiederm After the applications completes they like to pick up statics on the job, and getrusage is the traditional way to do that. 1142404571 M * Bertl is that something _you_ need from virtualization? 1142404629 M * eyck hmm 1142404639 M * ebiederm Bertl: yes. That is a routine piece. I can proably break the interface but not the functionality. 1142404641 M * eyck working getrusage would be nice.. 1142404664 Q * matta Ping timeout: 480 seconds 1142404671 M * Bertl ebiederm: I don't see any of the other parties use something like this right now, and actually I don't see much use for it 1142404688 M * Bertl ebiederm: doesn't mean that I'm against it ... 1142404733 M * Bertl ebiederm: just means for me that it doesn't warrant extra complexity 1142404763 M * Bertl s/warrant/justify/ 1142404801 M * ebiederm Bertl: I am pretty certain that the IBM guys have the same require burried somwhere. 1142404841 M * ebiederm Weird cases that most people don't care about but are important to some small group are why I want to keep the existing interfaces whenever possible. 1142404892 M * ebiederm But I don't think it is necessarily mandatory to keep the existing interfaces. 1142404964 M * ebiederm I need to look carefully, but I think in a clean implementation the part that would account for multiple pids per process would be 100 lines of code or less. 1142404995 M * Bertl anyway, I'm too tired for this right now ... so I will head off to bed ... 1142405010 M * ebiederm Sure. I don't mean to push. 1142405022 M * Bertl thanks for pointing out your intentions and requirements ... 1142405025 M * ebiederm Hopefully Kirill can live with not having all of his pids always mapped. 1142405043 M * ebiederm sleep well. 1142405050 M * Bertl have a nice day .. cya 1142405059 M * Bertl night folks, off now ... 1142405064 N * Bertl Bertl_zZ 1142405100 Q * ebiederm Quit: Leaving 1142406946 J * Aiken__ ~james@tooax6-080.dialup.optusnet.com.au 1142407269 Q * Aiken_ Ping timeout: 480 seconds 1142407549 J * coocoon ~coocoon@p54A06472.dip.t-dialin.net 1142407558 M * coocoon morning 1142407576 M * DataCompBoy YAW, coocoon! 1142407609 M * coocoon has it worked with dummy 1142407638 M * DataCompBoy http://linux-vserver.org/HowtoPrivateNetworking 1142407694 M * DataCompBoy working!:) 1142407879 Q * wasser Quit: Serverwechsel 1142407902 J * wasser ~wasser@ip86.ipax.at 1142407921 M * coocoon DataCompBoy: looks cool for that i have looked month ago, it looks easy, why have they written that it doesn't work 1142408025 M * DataCompBoy dunno -- i have decided to put that on wiki 1142408035 M * DataCompBoy also, I have found solution for one ssh to all vservers 1142408045 M * coocoon reall 1142408047 M * coocoon < 1142408048 M * coocoon y< 1142408050 M * DataCompBoy http://linux-vserver.org/HowtoSSHLogin 1142408054 J * click_ click@ti511110a080-3151.bb.online.no 1142408062 M * DataCompBoy YAW, click_! 1142408063 Q * click Read error: Connection reset by peer 1142408307 M * DataCompBoy i have contacted him for more detailed info -- will write complete page once got response 1142408482 M * coocoon DataCompBoy:oh cool, so i think it will also work for other services, i think right!? 1142408506 M * DataCompBoy looks so -- any that use login to auth 1142408564 M * coocoon DataCompBoy: coool nice to meet u ;-) 1142408570 M * DataCompBoy :)) 1142408578 M * DataCompBoy sutra of googling 4ever :) 1142408625 M * coocoon in native language u have more chance to find something 1142408626 M * DataCompBoy glad to be useful :D 1142408668 M * DataCompBoy in other words: DAO of googling take care about sutra googling followers 1142408739 M * DataCompBoy there in russia was several internet-searching contests, sponsored by yandex 1142408755 M * coocoon oh u have been there 1142408783 M * DataCompBoy no, I have found no time for... but several friends was. they are mosters! 1142408824 M * DataCompBoy oo... www.searchcontest.com/ -- yandex have sponsored contests at USA too 1142409026 M * coocoon i think u had won if u have been there ;-) 1142409047 M * DataCompBoy no -- winners are more mature... I have seen that -- that's right! 1142409103 M * coocoon had been there right 1142409335 Q * Aiken__ Remote host closed the connection 1142409343 J * Aiken ~james@tooax6-080.dialup.optusnet.com.au 1142409462 M * DataCompBoy YAW, Aiken! 1142411030 Q * Hmmmm Quit: Ex-Chat 1142411816 Q * shedi Quit: Leaving 1142413108 J * DataCompBoy2 ~datacompb@217.8.236.1 1142413394 Q * DataCompBoy Ping timeout: 480 seconds 1142414494 N * DataCompBoy2 DataCompBoy 1142414750 Q * `DoM` Quit: Byez 1142414866 J * shedi ~siggi@tolvudeild-200.lhi.is 1142414878 M * DataCompBoy Welcome, shedi! 1142415078 Q * Hollow Read error: Operation timed out 1142415650 J * Hollow ~hollow@home.xnull.de 1142416061 M * DataCompBoy Welcome, Hollow! 1142417016 M * shedi thank you DataCompBoy 1142417405 Q * coocoon Ping timeout: 480 seconds 1142418155 J * cryo ~say@217.146.242.85 1142418202 J * coocoon ~coocoon@p54A06472.dip.t-dialin.net 1142418402 M * DataCompBoy welcome cryo! 1142418405 M * DataCompBoy welcome back coocoon! 1142418536 M * coocoon hello 1142418574 M * SiD3WiNDR DataCompBoy: taking over Bertl_zZ 's job when he's zz? ;) 1142418689 M * coocoon very british 1142418842 M * DataCompBoy SiD3WiNDR: yup %) 1142418847 M * DataCompBoy :) 1142418865 M * DataCompBoy looks like we have about 12 hours difference, so can provide CM service:) 1142418967 M * DataCompBoy but... may be put on bot? :) 1142419016 M * DataCompBoy who will provide !rtfm commands, like on freenode #postgresql ? 1142419022 M * DataCompBoy very useful 1142419023 M * DataCompBoy :] 1142419834 M * coocoon SiD3WiNDR: do u have knowledge about the dev structures of vservers 1142419991 Q * rs Quit: rs 1142420111 M * SiD3WiNDR dunno what you mean 1142420225 M * coocoon SiD3WiNDR: which dev are needed, and why there must be special dev by using slackware 1142420249 M * SiD3WiNDR no idea 1142420284 M * phreak`` Bertl_zZ: bah, sorry :P 1142420296 M * coocoon ok 1142420299 M * coocoon thanx 1142420533 M * coocoon phreak``: u have slackware knowledge 1142420950 M * JimmyGulp whats this special dev thing? (I've got a test slackware server running redhat 7.3 legascy vservers 1142421075 M * coocoon JimmyGulp: i have had problems to run pkgtool in slack10.x guest, with normal devices after building vserver i got segment errors 1142421293 M * JimmyGulp ooh, I have a slackware guest on here too :) I'd forgotten about that 1142421325 M * coocoon JimmyGulp: i have got it to work, i wanted to know why it is different 1142421365 M * JimmyGulp what dev did you have to create? 1142421437 M * coocoon :-) 1142421442 M * coocoon JimmyGulp: I have found a working tamplate http://pastebin.com/603275 1142421483 J * lilalinux ~plasma@dslb-084-058-198-044.pools.arcor-ip.net 1142421493 M * coocoon JimmyGulp: i haven' t tested if i need fd or vc but after that the segments errors has been away 1142421611 M * JimmyGulp hmm, (running 2.6kernel, with util-vserver 0.30.210) I only have full, null, ptmx, random, tty, random, zero, and a pts directory 1142421632 J * matta ~matta@c-68-32-202-140.hsd1.pa.comcast.net 1142421654 M * coocoon JimmyGulp: me too, u have a slackware10.X guest 1142421658 M * coocoon running 1142421714 M * JimmyGulp hmm, thats a pretty seg fault... :/ 1142421763 M * coocoon JimmyGulp: u got it too 1142421802 M * JimmyGulp can you do an ls -la on /dev, see if I can work out what fixes it. 1142421830 M * coocoon in the guest 1142421847 M * JimmyGulp yeah 1142421881 M * coocoon JimmyGulp: http://pastebin.com/603280 1142421988 Q * Aiken Remote host closed the connection 1142422274 J * TrueLight ~kvirc@truelight.xs4all.nl 1142422302 M * TrueLight Hi all! 1142422323 M * TrueLight I had a question: normally when I press ctrl+c in a vserver, it stays in the vserver... now with the latest version, it goes back to the main system, with: 1142422344 M * TrueLight vlogin: select(): Interrupted system call 1142422353 M * TrueLight pretty annoying fact, and I have no idea what caused it 1142422400 M * TrueLight (util-vserver 0.30.210 btw, -r5 release of Gentoo) 1142422479 J * rs ~rs@office.dailymotion.com 1142422572 M * coocoon TrueLight: for this are bertl or daniel_hozac the right contact 1142422627 M * TrueLight coocoon: tnx, so I go sit here and wait for them to show up :) 1142422665 M * DataCompBoy TrueLight: downgrade to 0.30.209 :) 1142422674 M * DataCompBoy for me them working pretty fine 1142422685 M * daniel_hozac TrueLight: whoops, sorry! 1142422689 M * TrueLight DataCompBoy: it worked up till -r1 from 0.30.210 too :) 1142422694 M * daniel_hozac TrueLight: you can revert to -r4. 1142422704 M * TrueLight so it is a known problem, or? 1142422721 M * daniel_hozac nope, but vlogin isn't very well tested. 1142422744 M * TrueLight I now realise it somehow installed unstable versions... 1142422750 M * TrueLight Oops :) 1142422802 M * TrueLight but okay, will downgrade for now 1142422802 M * TrueLight tnx 1142422821 M * daniel_hozac could you perhaps test a patch for fixing this? 1142422835 M * TrueLight sure 1142422877 M * TrueLight (I am currently wondering why vconfig and util-vserver were running on non-stable versions.. :s) 1142422901 M * DataCompBoy why not? 1142422914 M * DataCompBoy :) non-stable not everytime mean "non-working" 1142422932 M * TrueLight yeah, but on production servers it isn't the best thing to do 1142422938 M * TrueLight unmasking a certain revision, acceptable 1142422939 M * DataCompBoy of course 1142422940 M * TrueLight but... 1142422950 M * TrueLight I should get myself checked! 1142422952 M * DataCompBoy ehh... i want to have copy-on-write feature 1142422972 Q * matta Read error: Operation timed out 1142422980 M * DataCompBoy but not sure is 2.1.0 stable for my purposes. :) 1142422980 M * TrueLight (I guess it comes from the time vserver was masked on amd64.. what ever) 1142423139 M * JimmyGulp coocoon, where did you get the vc directory stuff from? 1142423163 M * coocoon JimmyGulp: from 10.1 template 1142423189 M * JimmyGulp odd... my host doesn't have a /dev/vc, though it does have a /dev/vcc 1142423234 M * coocoon JimmyGulp: u mean it only works with the vc devs 1142423286 M * daniel_hozac TrueLight: http://daniel.hozac.com/vserver/util-vserver-0.30.210-vlogin.patch 1142423417 M * TrueLight daniel_hozac: k, give me some secs to compile it 1142423430 M * daniel_hozac TrueLight: remember to back out the previous vlogin patch first. 1142423444 M * TrueLight hehe, smart thing to say :) 1142423514 M * coocoon JimmyGulp: http://www.fumanchu.com/vserver 1142423746 M * TrueLight daniel_hozac: seems to work 1142423775 M * daniel_hozac TrueLight: meaning SIGINT doesn't kill it and goes to the bash in the vserver? 1142423781 M * TrueLight yup 1142423787 M * daniel_hozac cool, thanks. 1142423796 M * TrueLight yeah, works in multiple situations :) 1142423799 M * TrueLight no, thank you :) 1142423909 M * daniel_hozac Hollow, phreak``: new vlogin patch fixing SIGINT handling, http://daniel.hozac.com/vserver/util-vserver-0.30.210-vlogin.patch 1142424125 M * JimmyGulp coocoon, I'm not sure why having vc fixes it for you... If I use gdb, its the dialog app thats breaking, and on my machine, its dieing when it tries to access the mouse stuff (running without a mouse here) 1142424174 M * coocoon u have got my link 1142424180 M * coocoon JimmyGulp 1142424201 M * JimmyGulp yeah 1142424260 M * JimmyGulp just downloading now 1142424290 M * coocoon JimmyGulp: pkgtool remove gpm I mean, this is the mouse server 1142424439 M * coocoon daniel_hozac: u have maybe an answer why it works from slack10.X (guests) with such dev http://pastebin.com/603275, and with standard devs in created devs folder it gives out segs errors? http://pastebin.com/603280 1142424515 M * daniel_hozac where are the errors? 1142424535 M * daniel_hozac and isn't /dev/vc the same as /dev/pts, only in devfs speak? 1142424574 M * coocoon daniel_hozac: in the guest, starting pkgtool, ah cool sounds good 1142424588 M * daniel_hozac Hollow, phreak``: whoops, missed a few lines. updated it. 1142426660 M * DataCompBoy do I need to have quotas enabled on host, to make quotas work for guest? 1142426796 P * TrueLight 1142426854 M * coocoon DataCompBoy: thats my favour http://linux-vserver.org/Disk+Limits 1142426889 M * DataCompBoy ok, thanks :) 1142426897 M * DataCompBoy where is list of ALL pages on that wiki? 1142426899 M * DataCompBoy categorized 1142426939 M * coocoon http://linux-vserver.org/Linux-VServer 1142426951 M * coocoon http://linux-vserver.org/Documentation 1142427015 M * DataCompBoy that not all, only listed :) 1142427585 Q * brc Ping timeout: 480 seconds 1142427904 M * daniel_hozac http://linux-vserver.org/?action=find&find= 1142427906 M * daniel_hozac perhaps. 1142427984 M * daniel_hozac also, disk limits are not the same as quotas. disk limits will limit the entire vserver, while quotas will limit a specific user/group. 1142428089 Q * sladen Ping timeout: 480 seconds 1142428128 J * sladen paul@starsky.19inch.net 1142428628 M * coocoon daniel_hozac: right thanx i must learn a lot 1142428656 J * matta ~matta@c-68-32-239-173.hsd1.pa.comcast.net 1142428804 M * DataCompBoy hmmm.... oo there also very intresting PAM module: 1142428811 M * DataCompBoy pam_eps: A PAM module for ssh authentication against a remote server. 1142428824 M * DataCompBoy may be it also useful to have one ssh for every hosts? 1142428842 M * mnemoc i have one dropbear per host 1142428847 M * mnemoc guest i mean 1142428860 M * mnemoc of course also another on the host 1142428980 M * daniel_hozac DataCompBoy: if you're going to be running all of your services on the host, what's the point of using vserver at all? 1142429015 J * doener ~doener@i5387E549.versanet.de 1142429323 M * DataCompBoy users will log into vserver 1142429339 M * DataCompBoy thir sites works independed from main 1142429343 M * DataCompBoy and have no access to main 1142429492 A * DataCompBoy thinks ... or you right, and set up its own mail and spam checkers at ever vsite ... 1142430008 M * DataCompBoy o, one more intresting question. how can I limit size for only own files of vmachine, if I have vunified them? 1142430126 M * DataCompBoy  1142430150 M * DataCompBoy I have unable to set tagxid option to my partition...: 1142430172 M * DataCompBoy http://pastebin.com/603395 1142430204 M * coocoon ist it lvm or raid 1142430227 M * coocoon it is not recommended to use it on / 1142430228 M * DataCompBoy its ext3 1142430233 M * doener DataCompBoy: you cannot specify tagxid on remount 1142430247 M * doener and having tagxid on your rootfs is a bad thing 1142430269 M * DataCompBoy hmm... but I have only one partition for all purposes. exception is only /boot separate 1142430273 M * doener the tools touch various files in the startup sequence and those would get tagged 1142430294 M * coocoon u must resize rootfs 1142430317 M * doener time for some repartitioning then :) 1142430332 M * DataCompBoy =) 1142430359 M * coocoon here u find all vserver.org is a cool thing, http://linux-vserver.org/Remote+Server+Repartitioning 1142430367 M * coocoon thanx to all who has created it 1142430369 M * DataCompBoy then, i'll call to noc so they swap my drives... pffr 1142430392 M * DataCompBoy he he he, I know how to do that remote -- I prefer to do that from other system :D 1142431362 J * lilalinux_ ~plasma@dslb-084-058-233-007.pools.arcor-ip.net 1142431550 M * Hollow daniel_hozac: what did you change for SIGINT? 1142431678 N * click_ click 1142431735 Q * lilalinux Ping timeout: 480 seconds 1142432199 J * lilalinux__ ~plasma@dslb-084-058-214-099.pools.arcor-ip.net 1142432264 M * JimmyGulp would someone mind explaining what /etc/vservers//run should be? does it contain some text, or be a symbolic link to something else? 1142432433 M * Hollow it is a symlink to /var/run/vservers/ 1142432441 M * daniel_hozac Hollow: i messed up when i moved to Eselect over the while you had in there. 1142432482 M * Hollow hm, phreak just told me hitting ctrl+c twice terminates the vlogin process, does this fix it? 1142432505 M * daniel_hozac hitting it once should suffice. 1142432509 M * daniel_hozac and it's supposed to. 1142432519 M * daniel_hozac (TrueLight verified) 1142432524 M * Hollow well... no, ctrl+c in bash does not terminate the bash process 1142432548 Q * lilalinux_ Ping timeout: 480 seconds 1142432567 M * Hollow phreak``: ping.. 1142432567 M * daniel_hozac but due to the error in my patch, ctrl+c would interrupt select, and then vlogin would bail out. 1142432577 M * Hollow ah, and now it is fixed 1142432580 M * daniel_hozac right. 1142432583 M * Hollow ok, great 1142432587 M * coocoon DataCompBoy: run= Points to a file which will contain the XID of the running vserver. When the vserver is stopped, this can be a dangling symlink. 1142432601 M * coocoon DataCompBoy: http://www.nongnu.org/util-vserver/doc/conf/configuration.html 1142432687 M * coocoon JimmyGulp: sorry was for u 1142432694 M * Hollow phreak``: unping 1142432752 M * JimmyGulp coocoon, thanks 1142432842 M * coocoon JimmyGulp: uw, the image is good isn't it 1142432873 M * JimmyGulp yes, I like operas ability to turn css off ;) 1142432880 M * DataCompBoy oops, power failure here :) i 1142432887 M * DataCompBoy 'm shutting down 1142432909 M * DataCompBoy see you, all! 1142432909 Q * DataCompBoy Quit: Õðÿï 1142432916 M * coocoon JimmyGulp: sorry i mean template 1142432935 M * coocoon DataCompBoy: bye 1142433744 N * Bertl_zZ Bertl 1142433748 M * Bertl morning folks! 1142433769 M * daniel_hozac morning! 1142433814 M * coocoon hello bertl 1142433899 M * phreak`` heya Bertl :) 1142433950 M * phreak`` Bertl: just shoot, I'll look into it when I get back (probably in the evening or tomorrow morning) 1142433965 M * Bertl hey phreak``! just wanted to know with what branch/kernel you tested the sendfile stuff and what exactly was required to make it work ... 1142433990 M * Bertl and most important, does 2.1.1-rc13 work for you now (2.6.15 and 2.6.16?) 1142435315 J * pzYsTorM schak@dslb-082-083-040-006.pools.arcor-ip.net 1142435367 M * pzYsTorM morning. got a question: amd64 3000+... and vserver... does it work?... a 64bit debian with vserver-engine? 1142435512 M * Bertl I don't know about the debian part, everything else works fine :) 1142435524 M * doener heh :) 1142435589 M * Bertl pzYsTorM: read: mandriva 64bit works like a charm with a 64bit vanilla vserver kernel 1142435608 M * doener pzYsTorM: as you seem to be running debian amd64, maybe you know what sources.list should look like? my packages (etch) don't match what packages.debian.org says... :/ 1142435621 M * coocoon what do u# think about that how to from DataCompBoy http://linux-vserver.org/HowtoPrivateNetworking 1142435684 M * Bertl well, I had a look at it yesterday ... it's basically what we tell people since a long time, although the dummy0 is there just for fun 1142435747 M * coocoon maybe the ip looks better than 1.2.3.4 ;-) 1142435797 M * pzYsTorM doener: sorry, i dont have a 64bit machine... perhaps soon. its only a plan, because firsthop.de suxXXXxx enormously :( 1142435823 M * pzYsTorM Bertl: okay thanks ;) 1142437896 M * doener Bertl: is there a way to simulate default arguments for macros with cpp? 1142437915 J * ebiederm ~eric@ebiederm.dsl.xmission.com 1142438007 M * Bertl doener: not that I know of, if I got that question right :) 1142438041 M * doener like you can do in c++ void foo(int bar=5); 1142438075 M * Bertl as you cannot have cpp macros which take different number of arguments 1142438088 M * ebiederm Bertl: vararg macros are possible. 1142438107 M * ebiederm C99 standardized them. 1142438108 M * Bertl vararg yes, but not for evaluation 1142438131 M * Bertl what I mean is, you cannot have: 1142438141 M * Bertl #define test(a) (a) 1142438147 M * Bertl #define test(a,b) (a+b) 1142438205 M * ebiederm Right. 1142438235 M * ebiederm And I don't think there is anything you can do with the vararg pieces to make it happen. 1142438245 M * ebiederm Although there are some very creative people in the world.' 1142438272 M * Bertl trust me, I really pushed the limits of cpp with the syscall stuff 1142438293 M * doener heh 1142438316 M * Bertl ebiederm: http://vserver.13thfloor.at/Stuff/SYSCALL/syscall.h (if you are interested :) 1142438502 M * ebiederm __sc_asm load seems to get about as close as you can get ;) 1142438565 J * doener_ ~doener@i5387F47D.versanet.de 1142438656 J * lilalinux_ ~plasma@dslb-084-058-239-253.pools.arcor-ip.net 1142438696 Q * doener Read error: Operation timed out 1142438803 J * geree ~return@we1-as7873.alshamil.net.ae 1142438974 J * meandtheshell ~markus@85-124-9-106.dynamic.xdsl-line.inode.at 1142439004 M * Bertl welcome geree! wb meandtheshell! 1142439021 M * meandtheshell hi bertl ;-) 1142439042 M * geree hi Bertl:D 1142439042 Q * lilalinux__ Ping timeout: 480 seconds 1142439356 J * lilalinux__ ~plasma@dslb-084-058-244-050.pools.arcor-ip.net 1142439507 M * ebiederm Bertl: Anyway it felt like we didn't quite finish our conversation last night. Is there anything more in that general area we need to discuss? 1142439552 J * Smutje_ ~Smutje@xdsl-87-78-16-46.netcologne.de 1142439659 Q * Smutje Ping timeout: 480 seconds 1142439659 N * Smutje_ Smutje 1142439702 Q * lilalinux_ Ping timeout: 480 seconds 1142439826 Q * shedi Quit: Leaving 1142439873 M * Bertl ebiederm: probably, but I have to leave now .. I'm already late 1142439888 M * Bertl ebiederm: but I will be back in a few hours, maybe we can continue then 1142439900 M * Bertl okay, off for now ... back later 1142439905 N * Bertl Bertl_oO 1142439909 M * ebiederm Ok. I will check back later. 1142439911 Q * ebiederm Quit: Leaving 1142440107 M * Wonka linux-image-2.6-vserver-686 2.6.15+2.6.16-rc5-0experimental.1 just appeared in debian experimental... 1142440128 M * Wonka and several other vserver-images also 1142440316 J * stefani ~stefani@superquan.apl.washington.edu 1142440531 J * Loki|muh loki@satanix.de 1142441042 M * coocoon ok good bye 1142441044 Q * coocoon Quit: KVIrc 3.2.0 'Realia' 1142442081 J * lilalinux_ ~plasma@dslb-084-058-245-197.pools.arcor-ip.net 1142442438 Q * lilalinux__ Ping timeout: 480 seconds 1142442578 J * brc_ bruce@20151198033.user.veloxzone.com.br 1142442771 J * sachu ~sachu@220.226.48.206 1142443025 J * Viper0482 ~Viper0482@p54976ADF.dip.t-dialin.net 1142443518 J * shedi ~siggi@inferno.lhi.is 1142444060 Q * gerrit Quit: Client exiting 1142447487 Q * Viper0482 Remote host closed the connection 1142447652 J * Viper0482 ~Viper0482@p54976ADF.dip.t-dialin.net 1142448129 J * ntrs ~ntrs@68-188-37-15.dhcp.stls.mo.charter.com 1142448235 Q * sachu Quit: Leaving 1142448532 M * pzYsTorM if i want to set rss hardlimits at vs2.0.2-rc12... do i have to use /etc/vserver/bla/rlimits/rss.hard or /etc/vserver/bla/rlimits/rss or /etc/vserver/bla/rlimit/rss or /etc/vserver/bla/rlimit/rss.hard ? :/ 1142449048 J * bonbons ~bonbons@83.222.39.180 1142449148 M * daniel_hozac rlimits/rss 1142449163 M * daniel_hozac rlimits/rss.hard would work too, though. 1142449560 M * pzYsTorM hm... doesnt work... 1142449562 M * pzYsTorM kara:/# ls -l /etc/vservers/v3/rlimits 1142449562 M * pzYsTorM insgesamt 8 1142449562 M * pzYsTorM -rw-r--r-- 1 root root 6 2006-03-15 19:56 rss 1142449562 M * pzYsTorM -rw-r--r-- 1 root root 6 2006-03-15 19:56 vsz 1142449562 M * pzYsTorM kara:/# cat /etc/vservers/v3/rlimits/rss 1142449562 M * pzYsTorM 20000 1142449564 M * pzYsTorM kara:/# cat /etc/vservers/v3/rlimits/vsz 1142449564 M * pzYsTorM 30000 1142449574 M * pzYsTorM and we could alloc 250MB inside the guest 1142449728 M * pzYsTorM (pagesize is 4kb) 1142450061 Q * meandtheshell Quit: bye bye ... 1142450082 M * daniel_hozac what does /proc/virtual//limit 1142450086 M * daniel_hozac +say? 1142450118 M * pzYsTorM # cat /proc/virtual/1003/limit 1142450118 M * pzYsTorM PROC: 84 95 -1 0 1142450118 M * pzYsTorM VM: 172674 234716 -1 0 1142450118 M * pzYsTorM VML: 0 0 -1 0 1142450118 M * pzYsTorM RSS: 5721 8409 20000 0 1142450118 M * pzYsTorM ANON: 5159 5159 -1 0 1142450120 M * pzYsTorM FILES: 455 583 -1 0 1142450120 M * pzYsTorM OFD: 376 376 -1 0 1142450122 M * pzYsTorM LOCKS: 2 7 -1 0 1142450122 M * pzYsTorM SOCK: 22 22 -1 0 1142450124 M * pzYsTorM MSGQ: 0 0 -1 0 1142450124 M * pzYsTorM SHM: 0 0 -1 0 1142450149 M * pzYsTorM why does the VM limit doesnt work? 1142450152 M * daniel_hozac you sure you were doing the test in the right vserver? 1142450168 M * pzYsTorM yes, of course 1142450176 M * daniel_hozac it's as, not vsz. 1142450192 M * pzYsTorM the filename is "as" ? 1142450196 M * daniel_hozac yes. 1142450198 M * pzYsTorM oh ok! 1142450242 M * daniel_hozac as you can see, you haven't used more than 8409 pages of RSS. 1142450288 M * pzYsTorM VM: 28363 29296 30000 5 1142450288 M * pzYsTorM VML: 0 0 -1 0 1142450288 M * pzYsTorM RSS: 2429 2861 20000 0 1142450294 M * pzYsTorM thats it. thanks 1142450310 M * daniel_hozac you're welcome! 1142451324 Q * sladen Ping timeout: 480 seconds 1142451443 J * sladen paul@starsky.19inch.net 1142451633 Q * Smutje Quit: leaving 1142452687 J * Tesla weasel@tor.noreply.org 1142453569 Q * geree Quit: epic 1142454663 N * Bertl_oO Bertl 1142454668 M * Bertl evening folks! 1142454754 Q * shedi Quit: Leaving 1142455599 M * eyck evening 1142455782 J * shedi ~siggi@inferno.lhi.is 1142456239 Q * samuel_ Ping timeout: 480 seconds 1142456277 P * stefani I'm Parting (the water) 1142456299 M * pzYsTorM got a bug: 1142456300 M * pzYsTorM SHM: 0 -2548 -1 0 1142456309 M * pzYsTorM negative pagecount for shared memory? 1142456312 M * Bertl eek, yep that should not happen 1142456319 M * pzYsTorM 2.0.2-rc12 1142456324 M * Bertl how did you get it? 1142456337 M * pzYsTorM only started a guest 1142456358 M * Bertl is it reproduceable? 1142456362 M * pzYsTorM every time 1142456373 M * Bertl very good, how large is the guest? 1142456379 M * pzYsTorM 15GB 1142456386 M * pzYsTorM if u mean the FS 1142456399 M * Bertl hmm, probably not too easy to tar up then :) 1142456422 M * Bertl do you want to debug this? 1142456461 M * pzYsTorM yes, if its not to difficult and if you tell me where to start 1142456470 M * Bertl would require a few things, mainly enabling vserver debugging and tracing a run if possible without other guests running 1142456532 M * Bertl CONFIG_VSERVER_DEBUG=y 1142456549 M * Bertl CONFIG_DEBUG_BUGVERBOSE=y 1142456554 M * Bertl CONFIG_DEBUG_INFO=y 1142456559 M * Bertl CONFIG_DEBUG_KERNEL=y 1142456568 M * pzYsTorM oh... thats bad... because its a "sold" guest... he pays monthly... every reboot is bad... and every "intrusion" into the guest makes me guilty ;) 1142456594 M * Bertl well, I would say it should suffice to copy the guest somewhere, maybe even offline 1142456616 M * Bertl and then you can start and stop as often as you like 1142456627 M * pzYsTorM okay. that would be no problem. i have an unused partition and an unused IP 1142456699 M * pzYsTorM i come back in a few days, when kernel-vserver-debugging is enabled, the partition copied, and my brain is free for debugging ;) 1142456711 M * Bertl okay, great! 1142456752 M * Bertl will look into the relevant code sections, maybe I see something obvious ... 1142457256 M * doener_ pzYsTorM: 2.6.15? 1142457374 M * pzYsTorM .6 1142457375 M * pzYsTorM yes 1142457636 J * scott ~scott@dkhosting.net 1142457722 M * Bertl welcome scott! 1142457874 M * doener_ Bertl: the vx_ipcshm_* are used for shared mem usage tracking, right? 1142457901 Q * Tesla Quit: Nicola Tesla alive 1142457921 M * Bertl doener_: should be, IIRC 1142457929 M * doener_ ok 1142457940 A * doener_ forgot most stuff he knew about the vserver code :/ 1142457964 M * Bertl ah, well, don't worry, I have to check to code too ... 1142458348 J * coocoon ~coocoon@p54A0684D.dip.t-dialin.net 1142458358 M * coocoon hello 1142458361 M * daniel_hozac hi 1142458501 M * coocoon need a command for reading all text/ conf/ /bash files with output, to look for special words, can someone help 1142458502 M * coocoon me 1142458502 Q * Viper0482 Read error: Connection reset by peer 1142458552 M * daniel_hozac huh? 1142458574 M * mugwump coocoon: find /etc -type f | xargs grep -i 'word' 1142458583 M * coocoon ok cool 1142458595 M * coocoon thanx a lot 1142458597 M * mugwump that will search your config files (in /etc) for instances of "word" 1142458609 M * coocoon thats what i am looking for 1142458630 M * mugwump let me guess, changing host name / ip / etc? :) 1142458642 M * eyck grep -r word /etc/ 1142458656 M * mugwump eyck: smart alec :) 1142458662 M * daniel_hozac -H might be handy, too. 1142458668 M * coocoon very nice 1142458721 M * Bertl and -s :) 1142458842 M * coocoon daniel_hozac: it takes a long time for understand that but now I have it 1142458845 M * coocoon ;-) 1142460212 Q * bonbons Quit: Leaving 1142460399 Q * kilian Ping timeout: 480 seconds 1142460507 M * eyck who's alec? 1142460618 J * kilian kk@projects.verfaction.de 1142460696 M * coocoon mugwump: sorry have read it but now, no for removing visas 1142460794 M * mugwump eyck: ok, I spelt it wrong. "smart aleck" is in the dictionary :) 1142460903 M * Bertl hmm, not even google is working for me right now .. is there something loose on the internet? a new virus or so? 1142460918 M * eyck OMG, you broke the internet! 1142460922 M * eyck ...again 1142460965 M * daniel_hozac someone said some teleglobe routers broke. 1142461066 M * coocoon http://www.searchcontest.com/en/index.html 1142461140 M * coocoon they are learning 1142461195 M * Bertl okay, I take it as sign not to work this evening :) 1142461212 M * Bertl off for tonight ... back later tomorrow ... 1142461227 M * coocoon bye 1142461228 N * Bertl Bertl_zZ 1142461816 M * eyck have fun 1142461959 Q * scott Quit: leaving 1142462320 M * coocoon vserver guest rename shows this message "15869 141 sin41 ? Ss 0:00 init [6]" 1142462330 M * coocoon -renam +restart 1142462392 M * coocoon also everytime a tty10 will be created which initscript could be responsible for that 1142462458 M * daniel_hozac what type of file is tty10? 1142462467 M * daniel_hozac and i guess a grep should tell you that. 1142462565 M * coocoon it looks like an error log file 1142462585 M * daniel_hozac so it's a file, and not a device? 1142462619 M * coocoon ok u r right 1142462642 M * coocoon it reports that the other tty's can't be found 1142462925 Q * lilalinux_ Remote host closed the connection 1142463184 M * coocoon daniel_hozac: cool thanx 1142463426 J * Aiken ~james@tooax8-040.dialup.optusnet.com.au 1142463452 J * samuel_ ~samuel@levinux.UQAR.UQUEBEC.CA 1142463468 A * samuel_ is away: (Auto-Away after 10 mins) [BX-MsgLog On] 1142463705 Q * mkhl Ping timeout: 480 seconds 1142465269 Q * rs Quit: rs 1142466112 J * mkhl ~mkhl@200-148-40-56.dsl.telesp.net.br