1152403217 Q * schimmi Ping timeout: 480 seconds 1152405464 M * matti Bertl: Night :) 1152408610 Q * virtuoso Ping timeout: 480 seconds 1152408815 J * virtuoso ~s0t0na@shisha.spb.ru 1152409984 N * Ben_ Ben_zZz 1152412786 J * yahn ~yahn@c-71-236-68-108.hsd1.pa.comcast.net 1152412842 M * yahn i'm sorry to bother you but i really need help i was using kwrite and a message came up and now i have to hold a key in for a second for it to display on any program does anyone know how to fix it I would really appreciate help 1152413058 J * Kurt ~kmw@216.135.25.185 1152413069 P * yahn Leaving 1152418051 J * Kurt2 kmw@evvlinlwt-nas-03-s396.cinergycom.net 1152418261 Q * qfire Quit: leaving 1152418419 Q * Kurt Ping timeout: 480 seconds 1152419755 J * doener ~doener@i577BB5E1.versanet.de 1152419762 Q * virtuoso Ping timeout: 480 seconds 1152420169 Q * doener_ Ping timeout: 480 seconds 1152425304 Q * cmatei Ping timeout: 480 seconds 1152426579 Q * Kurt2 Ping timeout: 480 seconds 1152426654 J * DreamerC_ ~dreamerc@59.112.26.85 1152427029 Q * DreamerC Ping timeout: 480 seconds 1152427187 J * DreamerC ~dreamerc@59.112.19.158 1152427579 Q * DreamerC_ Ping timeout: 480 seconds 1152428267 J * DreamerC_ ~dreamerc@59.112.26.194 1152428659 Q * DreamerC Ping timeout: 480 seconds 1152428676 J * dna ~naucki@p54BCD848.dip.t-dialin.net 1152429842 Q * DreamerC_ Quit: leaving 1152429878 J * DreamerC ~dreamerc@59.112.26.194 1152432446 J * schimmi ~sts@port-212-202-73-176.dynamic.qsc.de 1152433366 J * dna_ ~naucki@p54BCE20A.dip.t-dialin.net 1152433612 Q * dna Ping timeout: 480 seconds 1152433887 Q * dna_ Ping timeout: 480 seconds 1152434096 J * Smutje ~Smutje@xdsl-87-78-60-182.netcologne.de 1152434116 J * dna ~naucki@p54BCE20A.dip.t-dialin.net 1152434512 J * dna_ ~naucki@p54BCE134.dip.t-dialin.net 1152434887 Q * dna Ping timeout: 480 seconds 1152434917 J * Viper0482 ~Viper0482@p54977A2E.dip.t-dialin.net 1152435062 J * [1]qos ~qos@211.90.82.139 1152435147 Q * qos Read error: Connection reset by peer 1152435148 N * [1]qos qos 1152435151 Q * qos Quit: 1152435214 J * qos ~qos@211.90.82.139 1152435262 J * Smutje_ ~Smutje@xdsl-87-78-6-2.netcologne.de 1152435379 Q * Smutje Ping timeout: 480 seconds 1152435379 N * Smutje_ Smutje 1152436175 J * virtuoso ~s0t0na@shisha.spb.ru 1152437342 J * coocoon ~coocoon@84.160.127.163 1152437653 M * coocoon morning 1152437855 M * daniel_hozac morning. 1152437896 M * daniel_hozac Bertl_zZ: a bit late, but sure :) 1152438686 J * SunSmile ~user@217.8.236.1 1152438730 Q * coocoon Ping timeout: 480 seconds 1152438743 M * SunSmile Hi averyone!!!!))) 1152439147 M * daniel_hozac hello 1152439414 J * coocoon ~coocoon@84.160.115.139 1152439499 J * coocoon3 ~coocoon@84.160.115.139 1152439511 Q * coocoon3 Quit: 1152441937 Q * phreak`` Quit: leaving 1152442253 J * phreak`` ~phreak``@140.211.166.183 1152442699 J * phreak``_ ~phreak``@140.211.166.183 1152442731 Q * phreak``_ Quit: 1152442748 Q * phreak`` Quit: leaving 1152443145 J * joep ~Joep@c53753253.cable.wanadoo.nl 1152443152 M * joep Hi there! 1152443247 M * joep I am looking for sime help on klogd. Installed vserver yesterday on my FC5 box. Went smooth. However, klogd starts up and dies 1152443287 M * joep Found in the archive a few references to the issue but no solution for FC5. Any suggestions? 1152443288 M * daniel_hozac yep. klogd inside a guest doesn't really make sense. 1152443326 M * joep Hi Daniel, thanks for your stuff! Made installing painless. 1152443328 M * daniel_hozac either comment the start/stop of klogd in /etc/init.d/syslog, or give the guest the virtualized klog, or ignore it. 1152443355 M * daniel_hozac painless is just the way i like it :) 1152443366 M * joep I saw that about klogd in a vserver not making sense. But how stopping it from starting - as it were 1152443463 M * SunSmile sup? 1152443464 M * SunSmile ))) 1152443572 M * joep OK, simple. As you suggested I commented the klog out of /etc/init.d/syslog start and stop. 1152443592 M * Wonka joep: i am running debian in my vserver, deinstalled sysklogd and installed inetutils-syslogd 1152443609 M * Wonka joep: that one doesn't contain a klogd 1152443639 M * daniel_hozac Wonka: i thought Debian had separate init scripts for syslogd and klogd? 1152443662 M * joep I read that Debian splits klogd and syslog - FC5 does not :( 1152443666 M * Wonka i don't like manipulating scripts, they tend to be overwritten now and then 1152443686 M * Wonka joep: but syslogd depends on klogd 1152443692 Q * SunSmile Quit: Прощайте человеки 1152443704 M * daniel_hozac no, it doesn't. 1152443735 M * Wonka Depends: libc6 (>= 2.3.6-6), klogd | linux-kernel-log-daemon 1152443767 M * Wonka sysklogd 1.4.1-18 1152443824 M * Wonka Package inetutils-syslogd doesnt 1152443842 M * daniel_hozac the packages might, but you can still disable the service from starting, no? 1152443997 M * Wonka you would have to edit or remove the init script, which i dislike 1152444029 M * daniel_hozac can't you just remove the symlinks to it from the appropriate runlevel directories? 1152444045 M * daniel_hozac alternatively, switch the S to K? 1152444046 M * Wonka editing config files is ok for me, editing init scripts not, as they're usually overwritten without asking (cause they're not config files) 1152444089 M * Wonka and changing rc?.d symlinks will also be reverted sometimes 1152444097 M * joep FC5 starts both loggers from one script 1152444132 M * joep wonka: inetutils-syslogd is not available for FC5 (?) 1152444170 M * Wonka joep: i don't know if FC has such packages or how they're named - you need to look yourself 1152444229 M * joep wonka: no problem - already did that as the package was mentioned in the archives. Ploughed the net before I joined here :) 1152444429 J * mire ~mire@240-167-222-85.COOL.ADSL.VLine.Verat.NET 1152444620 J * matled ~matled@85.131.246.184 1152445486 Q * schimmi Read error: Connection reset by peer 1152445503 J * schimmi ~sts@port-212-202-73-176.dynamic.qsc.de 1152445920 N * Ben_zZz Ben_ 1152446826 J * pisco ~pampel@80.135.141.120 1152446924 P * pisco 1152447149 Q * [PUPPETS]Gonzo Ping timeout: 480 seconds 1152448368 J * pisco ~pampel@80.135.141.120 1152448794 J * phreak`` ~phreak``@140.211.166.183 1152448905 J * phreak``_ ~phreak``@140.211.166.183 1152449075 Q * phreak`` Quit: leaving 1152449121 J * phreak`` ~phreak``@140.211.166.183 1152449164 Q * phreak`` Quit: 1152449168 Q * phreak``_ Quit: 1152449203 J * phreak`` ~phreak``@140.211.166.183 1152449799 Q * mire Quit: Leaving 1152449825 J * shedi ~siggi@130.208.221.254 1152449870 Q * pisco Ping timeout: 480 seconds 1152450352 Q * michal` Ping timeout: 480 seconds 1152450670 J * pisco ~pampel@80.135.189.16 1152450739 J * michal` ~michal@www.rsbac.org 1152450940 Q * doener Quit: leaving 1152451081 J * doener ~doener@host.magicwars.de 1152451519 J * DreamerC_ ~dreamerc@59.112.28.148 1152451528 Q * DreamerC_ Quit: 1152451577 J * DreamerC_ ~dreamerc@59.112.28.148 1152451755 Q * pisco Ping timeout: 480 seconds 1152451910 Q * DreamerC Ping timeout: 480 seconds 1152452003 J * doener_ ~doener@host.magicwars.de 1152452004 Q * doener_ Quit: 1152452371 M * mnemoc .oO 1152452546 M * doener I just finally setup screen+irssi on a remote server and still have problems that irssi (or maybe screen?) displays pure garbage when I resize my xterm... tried /redraw, /window balance and C-a C-l to no avail... And while trying stuff, I messed up my bash alias and started new irssis instead of using the current one... :( 1152452559 J * pisco ~pampel@80.135.189.16 1152452561 Q * Viper0482 Quit: one day, i'll find this peer guy and then i'll reset his connection!! 1152452562 M * doener thus the joins/parts 1152452831 N * Bertl_zZ Bertl 1152452837 M * Bertl morning folks! 1152452863 M * daniel_hozac morning! 1152452864 M * doener hey Bertl 1152452868 M * Bertl doener: doesn't happen here, i.e. works quite fine 1152452910 M * doener hmpf... :( 1152452921 M * Bertl (screen version 4.00.02) 1152452951 M * doener worked fine here locally as well... maybe I've trimmed that vserver a little too much, terminfo stuff is there though 1152452973 M * doener (debian on both, local and remote) 1152452984 M * Bertl btw, what do you (doener, daniel_hozac) think of the lock/admin features? 1152452999 M * doener must have missed that, got a short overview? 1152453013 M * Bertl http://vserver.13thfloor.at/Experimental/delta-lock-feat01.diff 1152453018 M * Bertl http://vserver.13thfloor.at/Experimental/delta-admin-feat01.diff 1152453043 M * Bertl it's just an idea for further improvements in this direction# 1152453232 M * doener ok, lock means "you can't get in here", and admin? 1152453331 M * Bertl admin means you can change flags and properties 1152453345 M * Bertl (not complete yet, e.g. utsname is not covered) 1152453398 M * Bertl basically removing admin makes the guest immutable against config changes 1152453585 M * doener sounds useful 1152453626 M * doener remote bin/irssi: screen -S irssi -D -R `which irssi` -- local bash alias: alias irssi='ssh irc.magicwars.de -t ~/bin/irssi' 1152453630 M * doener do these look good? 1152453725 N * DreamerC_ DreamerC 1152453779 J * mire ~mire@240-167-222-85.COOL.ADSL.VLine.Verat.NET 1152453880 Q * pisco Ping timeout: 480 seconds 1152454033 M * Bertl I simply start screen (with bash) and then irssi 1152454072 M * Bertl auto resizes when I logon with a different sized terminal 1152454106 Q * mire Quit: Leaving 1152454113 M * Bertl doener: the idea of admin+lock is that the private namespace will become 'private' :) 1152454207 M * Bertl so, together with a per context loop device, cryptoloop could become an option 1152454275 M * doener that sounds nice :) 1152454276 M * daniel_hozac of course, if the host admin _really_ wanted to to enter the guests, he'd just comment those lines :) 1152454292 M * doener brb 1152454294 Q * doener Quit: leaving 1152454306 M * Bertl daniel_hozac: yes, of course, you can always modify the kernel 1152454393 M * Bertl I'm not aiming for a guarantee here just for a feature :) 1152454428 M * Bertl note that it should be possible to shutdown such a guest too 1152454450 M * Bertl btw, the testing showed 'interesting' results with the tools 1152454467 M * Bertl i.e. it might be worth to add a few more error checks to the mix :) 1152454497 M * daniel_hozac such as? 1152454528 J * Kurt kmw@evvlinlwt-nas-07-s87.cinergycom.net 1152454538 M * Bertl daniel_hozac: http://paste.linux-vserver.org/170 1152454543 M * Bertl welcome Kurt! 1152454573 M * Kurt hey, sexy 1152454573 M * Bertl daniel_hozac: ps is then executed on the host :) 1152454617 J * doener ~doener@host.magicwars.de 1152454636 M * Bertl wb doener! 1152454640 M * daniel_hozac Bertl: hmm, interesting. 1152454660 M * Bertl daniel_hozac: but maybe my tools are just outdated on that test host 1152454877 M * daniel_hozac Bertl: -17 is EEXIST, no? 1152454878 J * pisco ~pampel@80.135.189.16 1152454964 M * daniel_hozac Bertl: it catches that and tells chcontext to try migrating instead. 1152455006 M * Bertl hmm, right, let me check if I pasted the right one :) 1152455150 M * Bertl seems to be a kernel issue not tool issue, checking now 1152455361 M * Bertl hrm, I must have been very tired :) 1152455370 M * daniel_hozac hehe. 1152455381 M * Bertl (!ret) -> (ret) :) 1152455420 Q * pisco Ping timeout: 480 seconds 1152455480 M * daniel_hozac lol, and i missed that. 1152456183 M * Bertl okay, fixed that in place ... too embarassing for a separate patch :) 1152456256 J * pisco ~pampel@80.135.189.16 1152456523 M * Hollow Bertl: the values in /proc/virtual/*/{cacct,limit} are uint64_t right? 1152456535 M * Hollow because they get converted to unsigned long in the printf 1152456541 M * Hollow but the structs contain uint64_t.. 1152456546 M * Bertl they are .. special :) 1152456563 M * Hollow :) 1152456580 M * Bertl basically we handle them as uint64, but we don't want to see the ~0ULL 1152456603 M * Hollow yes.. that's how i came across it 1152456611 M * Bertl the internal values are arch specific, the userspace interface is uin64 1152456614 M * Bertl *uint64 1152456648 M * Bertl so in userspace it's the best to work with uint64 and handle ~0ULL special 1152456668 M * Bertl (or even better *LIM_INFINITY) 1152456669 M * Hollow special in the sense of N/A? 1152456689 M * Bertl no, the laste 3 values are well defined in the includes IIRC, sec 1152456750 M * Bertl #define CRLIM_UNSET (0ULL) 1152456751 M * Bertl #define CRLIM_INFINITY (~0ULL) 1152456751 M * Bertl #define CRLIM_KEEP (~1ULL) 1152456755 M * Hollow well, it's not about the limits, it's about the accounting 1152456775 M * Hollow i basically need to know the scanf format and max value 1152456779 M * Hollow to parse /proc 1152456790 M * Bertl why would you want to do that? 1152456816 M * Hollow to collect statistics? 1152456835 M * Bertl I mean, you can get the whole info, more precise and faster from the kernel interfaces? 1152456844 M * Hollow where do i get the accounting? 1152456848 M * Hollow i can just get the limits 1152456850 M * Bertl without generating virtual files and parsing stuff ... 1152456865 M * Bertl well, if you need/want that info, we add a kernel interface 1152456884 M * Hollow great, i need that for cacct, cvirt and limit 1152456884 M * Bertl I'm not really for parsing proc stuff, you know :) 1152456908 M * Bertl the limit should be readable already, IIRC 1152456923 M * Bertl cvirt is composed of vhi_name and other values 1152456966 M * Bertl cacct is only partially relevant, we have to look into the details (i.e. what makes sense to export/monitor) and we should move from ticks to msec 1152456987 M * Bertl we probably have a lot of other stuff we can expose to userspace 1152457011 M * Hollow for now we only collect the net stuff in cacct 1152457059 M * Bertl 'we' here means? 1152457087 M * Hollow the guy doing the stats collector and me 1152457095 M * Hollow he is only on freenode 1152457109 M * Bertl okay, so you basically need an interface to the socket accounting, yes? 1152457122 M * Hollow right 1152457140 M * Bertl okay, will look into it 1152457143 M * Hollow for the other things in cacct i have no idea what it is ;) 1152457201 M * Hollow maybe you can add some things of cvirt to vx_get_info, like biasuptime, the thread and loadavg 1152457306 Q * doener Quit: leaving 1152457340 J * doener ~doener@host.magicwars.de 1152457376 M * Bertl Hollow: wouldn't the 'uptime' make more sense there? 1152457382 M * Hollow Bertl: and i still don't see where i could get limit accounting 1152457388 M * Hollow Bertl: sure.. 1152457466 Q * qos Quit: HydraIRC -> http://www.hydrairc.com <- 100,000+ downloads can't be wrong 1152457515 M * Bertl Hollow: okay, expect a patch in the next few hours (for testing) 1152457536 M * Hollow tia! 1152457539 M * Bertl I just need to restructure a few things ... which also will need some testing 1152457553 M * Hollow sure, i can test whatever needs to be tested ;) 1152457558 M * Bertl (so basically I'll combine them) 1152457659 M * doener ok, resizing works now :) was a bug in aptitude that caused SIGWINCH to be blocked by nearly everything 1152457846 J * bonbons ~bonbons@83.222.39.166 1152458060 Q * pisco Ping timeout: 480 seconds 1152458685 Q * joep Quit: Leaving 1152458865 J * pisco ~pampel@80.135.189.16 1152459539 J * mire_ ~mire@240-167-222-85.COOL.ADSL.VLine.verat.net 1152459623 M * Hollow Bertl: what about the list of running context ids? should i get those from /proc or will there be an interface too? 1152459656 M * Bertl hmm, shouldn't your monitor already _know_ that? 1152459676 M * Bertl for a new start I'd suggest to scan the entire id range once 1152459681 M * Hollow currently we readdir() /proc/virtual and look for numbers 1152459801 J * cdrx ~legoater@cap31-3-82-227-199-249.fbx.proxad.net 1152459806 M * Bertl wb cdrx! 1152459855 M * cdrx hi bertl 1152459856 M * matti Bertl: ;-) 1152459862 M * Bertl hey matti! 1152459866 M * matti What's up? 1152459871 M * matti Hollow: ;-) 1152459886 M * Bertl matti: working on Linux-VServer improvements of course :) 1152459951 M * matti ;-) 1152459972 M * matti Bertl: BTW, when 2.0.2 will be available? 1152460000 M * Bertl ah, that#s easy ... when it's done! :) 1152460015 M * matti Hehehe. 1152460029 M * matti ;-) 1152460143 M * mnemoc like e17 :p 1152460327 M * Bertl daniel_hozac, doener: is the return 0 (unconditionally) in vc_set_sched_v2() intentional, or should we pass on the do_set_sched() return code 1152460335 M * Bertl (or doesn't it matter at all) 1152460517 M * daniel_hozac do_set_sched doesn't fail, does it? 1152460542 M * Bertl good point, so we pass it on for the future, right? 1152460555 M * matti ;p 1152460591 M * daniel_hozac yeah, seems like the right thing to do. 1152460730 Q * pisco Ping timeout: 480 seconds 1152461534 J * pisco ~pampel@80.135.189.16 1152461797 M * Bertl okay, of for dinner, back shortly ... 1152461806 N * Bertl Bertl_oO 1152462187 Q * DreamerC Remote host closed the connection 1152462711 J * DreamerC ~dreamerc@59.112.28.148 1152462751 M * derjohn Bertl_oO, pong .... (may a long RTT, though) 1152463019 Q * cdrx Quit: Leaving 1152463277 P * dusty http://www.smashthestack.org/ 1152463395 Q * pisco Ping timeout: 480 seconds 1152463693 M * Kurt a free, open-content wiki-based encyclopedia: http://opencycle.vacommunity.net 1152463948 N * Bertl_oO Bertl 1152463957 M * Bertl Kurt: like wikipedia? 1152463966 M * Kurt the same goal, yes 1152463978 M * Kurt but we aim to fix what we see as internal structural problems in Wikipedia 1152463982 M * Kurt that can never be repaired from within 1152464050 M * Bertl and that would be? 1152464067 M * Kurt mainly, a system which allows megalomaniacs and power-trippers to run amok 1152464079 M * Kurt who are more concerned with enforcing every last consequence of the most obscure corner of policy 1152464085 M * Kurt rather than contributing good content 1152464092 M * Kurt people who care more about WHY people are there than what they do there 1152464103 M * Kurt people who insist on worrying about things that DON'T ACTUALLY MATTER and DON'T ACTUALLY HURT ANYTHING 1152464142 M * Bertl aha, hmm .. sounds like fanatics :) 1152464178 M * derjohn Bertl, as I have to leave soon: you ping-ed me (I am honored ;)) ? 1152464200 M * Bertl derjohn: yeah, but I cannot remember why, actually :) 1152464251 J * pisco ~pampel@80.135.189.16 1152464264 M * derjohn Bertl, no problem, I often have the same problem. reason for this: my ID card, I guess they listed the wrong birth date ;:) 1152464653 Q * mire_ Quit: Leaving 1152464679 Q * kaner Remote host closed the connection 1152464691 J * kaner kaner@strace.org 1152465827 J * pusling pusling@195.215.29.124 1152465875 Q * pisco Ping timeout: 480 seconds 1152466214 J * [PUPPETS]Gonzo gonzo@langweiligneutral.deswahnsinns.de 1152466674 J * pisco ~pampel@80.135.189.16 1152467515 Q * pisco Ping timeout: 480 seconds 1152467527 J * s0undt3c1 ~s0undt3ch@bl7-250-78.dsl.telepac.pt 1152467982 Q * s0undt3ch Ping timeout: 480 seconds 1152467992 N * s0undt3c1 s0undt3ch 1152468330 J * pisco ~pampel@80.135.189.16 1152469895 Q * pisco Ping timeout: 480 seconds 1152470297 J * sod ~sod@fo02.uniques.ro 1152470456 M * sod helo 1152470460 M * daniel_hozac hi 1152470463 M * Bertl welcome sod! 1152470484 M * sod let me be a fool for a minute; we have a setup with 2.6.17-vs2.0.2-rc24 and util-vserver: 0.30.210 1152470490 M * sod actually we have more of them 1152470499 M * Bertl okay 1152470499 M * sod but I just ran into a stupid problem 1152470509 M * sod with only one of them 1152470513 J * pisco ~pampel@80.135.189.16 1152470517 M * sod I get no space left on device in vservers 1152470519 M * sod all of them 1152470532 M * sod but no dlimits are enforced 1152470562 M * sod wso03:/# df -h 1152470562 M * sod Filesystem Size Used Avail Use% Mounted on 1152470562 M * sod /dev/hdv1 73G 4.1G 65G 6% / 1152470562 M * sod none 16M 16M 0 100% /tmp 1152470562 M * sod wso03:/# dd if=/dev/zero of=/tmp/100M bs=1M count=100 1152470563 M * sod dd: writing `/tmp/100M': No space left on device 1152470582 M * daniel_hozac you see the /tmp entry in df, right? 1152470586 M * sod vserver root is on /vservers, on the same partition as / 1152470592 M * daniel_hozac it's a 16 MiB tmpfs by default. 1152470594 M * Bertl well, /tmp is a virtual (tempfs) filesystem of 16MB size 1152470610 M * sod yes, sorry 1152470612 M * sod bad example 1152470614 M * Bertl sod: so testing with /tmp is not the best choice :) 1152470629 M * sod it is the same with of=/100M 1152470649 M * daniel_hozac for just one guest or all of them on that host? 1152470666 M * sod all of them on the host 1152470684 M * sod it is somehow connected with the tmpfs, I think now 1152470692 M * sod it's about 16MB free space 1152470715 M * Bertl well of=/100M will not end up on tmpfs 1152470723 M * sod yes 1152470737 M * sod we never had this problem before 1152470751 M * Bertl which one now? 1152470772 M * Bertl tmp being 16M or dlimits not accounting/enforcing? 1152470802 M * daniel_hozac disk limits would show in df though. 1152470807 M * sod no, sorry, let's forget about /tmp 1152470811 M * sod it was a bad example 1152470818 M * sod bad habbit of using tmp 1152470837 M * sod the output is the same if I use of=/ 1152470846 M * sod and we enforce no dlimits on that machine 1152470866 M * Bertl okay, elt's get a few more infos, e.g. testme.sh output and context id/dlimit setup for that guest 1152470872 M * sod ok 1152470907 M * sod wso:/usr/src/util-vserver-0.30.210# sh testme.sh 1152470908 M * sod Linux-VServer Test [V0.15] Copyright (C) 2003-2006 H.Poetzl 1152470908 M * sod chcontext is working. 1152470908 M * sod chbind is working. 1152470908 M * sod Linux 2.6.17-vs2.0.2-rc24 #3 Mon Jul 3 05:12:07 PDT 2006 i686 1152470909 M * sod Ea 0.30.210 273/glibc (DSa) 1152470910 M * sod VCI: 0002:0001 273 03000076 (TbLgnP) 1152470912 M * sod --- 1152470916 M * sod [000]# succeeded. 1152470918 M * sod [001]# succeeded. 1152470920 M * sod [011]# succeeded. 1152470922 M * sod [031]# succeeded. 1152470924 M * sod [101]# succeeded. 1152470926 M * sod [102]# succeeded. 1152470928 M * sod [201]# succeeded. 1152470930 M * sod [202]# succeeded. 1152470932 M * sod ih, long paste, sorry 1152470936 M * daniel_hozac please use paste.linux-vserver.org for pastes. 1152470942 M * sod yes, sorry 1152470974 M * Bertl okay, what xid does the guest have? 1152470977 M * sod no dlimit setup 1152470995 M * Bertl hmm, so you expect it to work without setup? 1152471042 M * sod no, I don't want it to work, I don't want any limits 1152471070 M * Bertl ah, okay, but you feel you encounter a limit 1152471071 M * sod just that we get no space left on device errors in guests, although we do not use dlimit or anything else 1152471089 M * Bertl well, could you also do a df -i inside the guest? 1152471117 M * sod one line paste: /dev/hdv1 9633792 133576 9500216 2% / 1152471221 M * Bertl okay, that is fine, without limits set, I don't think you will get 'No space left on device' except for /tmp 1152471280 M * sod well, we do :) 1152471308 M * sod I know, it's strange, we never encountered this before 1152471310 M * Bertl unusual ... okay, let's investigate a little ... 1152471325 M * sod we're looking now in the underlying storage 1152471338 M * Bertl check the dmesg output first 1152471344 M * sod maybe the patches for the external storage somehow screw things up 1152471346 M * Bertl (maybe even upload it somewhere) 1152471349 M * sod dmesg is fine 1152471355 M * sod as in no errors 1152471359 M * Bertl okay 1152471396 J * pisc1 ~pampel@80.135.189.16 1152471483 M * daniel_hozac you need patches for external storage? 1152471535 Q * pisco Ping timeout: 480 seconds 1152471880 M * sod daniel_hozac: yes, it's a fibrenetix storage 1152472349 J * pisco ~pampel@80.135.189.16 1152472467 M * Bertl I'd say the storage device is _below_ the filesystem layer, no? 1152472939 Q * [PUPPETS]Gonzo Ping timeout: 480 seconds 1152473099 M * sod yes, but it could somehow cause no space left on device 1152473114 M * Bertl hmm, how? 1152473137 M * sod dunno, that's what we-re looking into right now 1152473141 M * sod we're 1152473216 M * Bertl let's upload the contents of /proc/virtual//* somewhere (e.g. paste.linux-vserver.org) 1152473279 M * Kurt you know, if Italy wins today, then the US should be named co-champion 1152473970 Q * pisco Ping timeout: 480 seconds 1152474694 Q * Kurt Ping timeout: 480 seconds 1152474785 J * pisco ~pampel@80.135.189.16 1152475624 J * sezuan matthias@agamemnon.ipv6.scheff32.de 1152475686 M * sezuan can give a vserver the permission to give its processess nice values below 0? 1152475719 M * Bertl yes, it's a capability 1152475732 M * sezuan Is this documented somewhere? 1152475765 M * Bertl grep NICE /usr/include/linux/capability.h 1152475769 M * Bertl #define CAP_SYS_NICE 23 1152475880 M * Bertl but also allows realtime scheduling (sidenote) 1152475915 M * sezuan that's no problem. 1152476069 M * sezuan well, I found the CAP_* stuff. But I try to find, how to use it. 1152476070 Q * shedi Read error: Connection reset by peer 1152476078 M * sezuan /etc/vserver//flags ? 1152476112 M * Bertl it's a bcapability 1152476114 M * tokkee What is anonymous memory? I keep forgetting what exactly that is :-( 1152476142 M * Bertl memory not backed by files or shared space 1152476156 M * sezuan what does that mean? there is BCaps field in /proc/virtual/49157/status. 1152476189 M * tokkee Bertl: thx. 1152476200 M * Bertl sezuan: that means that you want to set it in different file, and you should switch to static contexts :) 1152476232 M * Bertl sezuan: the wiki seems unreachable for me, so I cannot lookup the config name ATM 1152476262 M * sezuan too many documentation resources. 1152476290 M * sezuan ah yes, the page is anavailable. 1152476525 Q * pisco Ping timeout: 480 seconds 1152476539 M * sezuan Bertl: thanks, it works. 1152476630 Q * bonbons Quit: Leaving 1152476652 M * Bertl page is back now, btw :) 1152476917 M * sezuan I'll take a look. 1152476975 M * sezuan I there somewhere a more or less complete configuration description? 1152476982 J * oliwel ~mail-at-o@host-62-245-151-178.customer.m-online.net 1152476984 M * Bertl http://linux-vserver.org/alpha+util-vserver 1152476992 M * Bertl check out the flower page 1152476996 A * oliwel waves hello to the crowd 1152477002 M * Bertl welcome oliwel! 1152477008 M * sezuan the flower page seems good. 1152477023 M * oliwel Hi Bertl - nice to see you 1152477030 M * Bertl oliwel: yeah, LTNS 1152477037 M * oliwel havent been here for a while... 1152477062 J * shedi ~siggi@130.208.221.254 1152477070 M * Bertl wb shedi! 1152477328 J * pisco ~pampel@80.135.189.16 1152477333 M * tokkee Is unspec traffic in /proc/virtual//cacct the same as traffic going thru raw sockets? 1152477366 M * tokkee What are examples for "other" traffic? 1152477461 M * oliwel Be 1152477500 M * oliwel Bertl: just came to my mind....we discussed about syslog some time ago - you told me that sharing a socket file throug vserver guests is a bad idea-....why ? 1152477539 M * Bertl tokkee: check out include/linux/vs_socket.h 1152477574 M * tokkee Bertl: I will... thx. 1152477595 M * Bertl not sure we have any with type 0 (which might be a bug actually) 1152477615 M * Bertl will double check that shortly 1152477765 N * Ben_ Ben_zZz 1152477805 M * tokkee A'right. 1152477833 M * tokkee Is the correct spelling "Linux-VServer" or "Linux-Vserver" or still something else? ;-) 1152477847 M * Bertl the first one, see logo :) 1152477874 M * tokkee Okay... so I was doing things right ;-) 1152478177 M * tokkee Bertl: type 0 is unix traffic? 1152478705 M * Bertl include/linux/socket.h gives the families 1152478737 M * Bertl we currently map 1 and 2 to 1,2 1152478761 M * Bertl and 10->3 all else to 4 1152478791 M * Bertl I think we should also map 0->0 1152478842 M * Bertl and PF_PACKET might be interesting too 1152478855 M * Bertl tokkee: are you interested in testing a patch there? 1152478919 M * tokkee Bertl: If it's not likely to break my system, I can try it ;-) 1152479130 Q * pisco Ping timeout: 480 seconds 1152479356 Q * sezuan Quit: leaving 1152479530 M * Bertl tokkee: http://vserver.13thfloor.at/Experimental/delta-sock-feat01.diff 1152479617 M * tokkee Bertl: I will check that later. I'm currently running vs2.0.2-rc21 but I supose that patch should still apply... 1152479626 M * Bertl very likely 1152479775 M * tokkee Bertl: Is the problem with HTTPS traffic already solved? 1152479785 M * Bertl which one? 1152479844 M * tokkee HTTPS traffic did not appear in /proc/virtual//cacct... 1152479861 M * tokkee We were talking about that a couple of weeks ago... 1152479888 M * Bertl interesting, somehow missed that ... daniel_hozac? 1152479940 M * Bertl I would not see how https would use different sockets than http 1152479946 M * tokkee I think you (or maybe daniel_hozac) said it's caused by Apache using sendfile (or something like that)... 1152479958 M * Bertl ah, yes, sendfile _is_ different 1152479978 M * Bertl but sendfile means that https traffic _is_ accounted, while http (via sendfile) is not 1152480008 M * Bertl anyway, I have no idea if we addressed that yet or not 1152480046 J * Smutje_ ~Smutje@xdsl-87-78-85-82.netcologne.de 1152480071 M * tokkee Ah yes... it was the other way around ;-) 1152480074 M * Bertl tokkee: but if somebody tests it, we could add special accounting for sendfile traffic too 1152480088 M * Bertl (if we not already did so :) 1152480089 M * tokkee Bertl: I'd test that... 1152480128 M * Bertl okay, let's see how the socket patch goes, I'll prepare something ontop of that 1152480164 Q * Smutje Ping timeout: 480 seconds 1152480164 N * Smutje_ Smutje 1152480290 M * tokkee PF_PACKET is UDP traffic? 1152480314 M * Bertl it's everything using PF_PACKET sockets :) 1152480430 M * tokkee Ga... I was confused ;-) 1152480630 M * tokkee Bertl: Hmm... the patch did not apply... I will try to apply it manually... 1152480665 M * Bertl ah, you are with stable, this is against devel 1152480720 M * tokkee Yes... 1152481055 M * oliwel gonna catch some sleep....n8 guys 1152481059 Q * oliwel Quit: Chatzilla 0.9.74 [Firefox 1.5.0.4/2006050817] 1152481078 M * tokkee Tz... those German guys ;-) 1152481086 M * tokkee neight does not make much sense ;-) 1152481631 J * Aiken ~james@tooax6-142.dialup.optusnet.com.au 1152481636 M * insomniac hi there :) 1152481651 M * Bertl welcome Aiken! insomniac! 1152481657 M * insomniac hi Bertl :) 1152481692 M * insomniac that's a nice day ;) 1152481696 M * Aiken hello 1152481840 Q * pisc1 Ping timeout: 480 seconds 1152481970 M * tokkee Bertl: What does the second part of the patch for include/linux/vserver/cvirt_def.h do? 1152481997 M * tokkee Bertl: I cannot find any for-loop in my version... 1152482035 M * Bertl that's a debug output, you can safely ignore it 1152482058 M * tokkee Okay. 1152482071 Q * coocoon Read error: Connection reset by peer 1152482071 M * tokkee And I don't have cvirt_init.h and cvirt_proc.h... 1152482072 M * Aiken Bertl I am still using 2.6.17-vs2.1.1-rc24 on x86, have not had any problems 1152482234 M * Bertl good, what about the alpha? 1152482346 M * Aiken have not been using it much, I'll check what kernel it has shortly 1152482407 M * tokkee Bertl: Where do I find equivalents in vs2.0.2? 1152482462 J * coocoon ~coocoon@84.160.115.139 1152482529 M * Bertl welcome coocoon! 1152482701 M * Aiken the alpha is still 2.6.17-rc5.-vs2.1.1-rc21.3 1152482710 M * Aiken have not used it enough to stress test it 1152482727 M * coocoon hey bertl 1152482734 M * Aiken all I can say is for how much I have used the alpha it has run fine with one guest 1152483819 J * mire ~mire@240-167-222-85.COOL.ADSL.VLine.verat.net 1152484274 M * Bertl wb mire! 1152484274 Q * shedi Read error: Connection reset by peer 1152484645 Q * dna_ Quit: Verlassend 1152485277 J * shedi ~siggi@130.208.221.254 1152487070 Q * shedi Read error: Connection reset by peer 1152487722 Q * schimmi Ping timeout: 480 seconds 1152487955 J * shedi ~siggi@130.208.221.254 1152488454 M * Bertl daniel_hozac, doener: anybody still awake? :) 1152488600 M * mnemoc *g* 1152488625 M * Bertl they already know that work is coming when I ask this way :) 1152488675 M * mnemoc *G*