1162253147 J * ms_ ~ms@arkansas.doc.ic.ac.uk 1162253312 J * lilalinux_ ~plasma@dslb-084-058-220-007.pools.arcor-ip.net 1162253662 Q * kugg Ping timeout: 480 seconds 1162253740 Q * lilalinux Ping timeout: 480 seconds 1162253890 J * debugger_ ~Rui@217.129.151.190 1162254317 Q * debugger Ping timeout: 480 seconds 1162255287 Q * debugger_ Quit: Fui embora 1162257278 Q * lilalinux_ Remote host closed the connection 1162257507 J * s0undt3ch_ ~s0undt3ch@81.193.56.122 1162257940 Q * s0undt3ch Ping timeout: 480 seconds 1162257940 N * s0undt3ch_ s0undt3ch 1162258114 J * kugg kugg@illvilja.org 1162259633 M * Piet is it possible (and does it make sense) to have a vserver running on multiple partitions? 1162259672 M * Piet i'm wondering whethe rit makes any sense to try to 'emulate' a common üpartitioning scheme within a vserver by creating multiple partitions on the host system 1162259788 M * Bertl well, if it has administrative advantages, why not? 1162259822 M * Bertl i.e. think limited tmp space (tmpfs) or specific /home, /opt, etc 1162259829 M * Bertl maybe even shared between guests? 1162259844 M * Skram woops, wrong channel 1162259913 P * litage Leaving 1162259928 M * doener without saying anything? 1162259968 Q * s0undt3ch Ping timeout: 480 seconds 1162260082 M * cehteh hmpf 1162260086 M * cehteh computers sucj 1162260088 M * cehteh k 1162260160 M * Bertl doener: well, it _was_ the wrong channel to say that, after all :) 1162260191 J * s0undt3ch ~s0undt3ch@81.193.58.48 1162260492 M * doener heh :) 1162260526 M * Piet Thanks Bertl. I'm setting up a system where I'll only have a single vserver in it, which is supposed to get all resources the host doesn't neccesarily need. 1162260569 M * Piet Is it safe to allocate the very most HDD space to the vserver partition or should I reserve some amount for the host /home ? 1162260589 M * Piet well, rather /root than /home 1162260613 M * Piet I'm thining of situation s where I need to download a software package on the host and need to compile it there, too. 1162260658 M * Piet I think I cannot install software on the host system which is stored in the guest, right (despite of it being a diffenrent security context)? 1162260670 M * Bertl well, you can limit the guest too (dlimits) if you want to reserve space 1162260782 Q * s0undt3ch Quit: leaving 1162260790 J * s0undt3ch ~s0undt3ch@81.193.58.48 1162260811 M * Piet hmm, so you thinkj if i needed extra space on the host i could set / decrease the dlimit and use the free space on the partition mounted to /var/libvservers to store data the host will use? 1162260834 M * Piet /var/lib/vservers 1162260868 M * Piet hmm, no i don't think this is what you meant 1162260882 M * Piet it wouldn't make much sense. 1162260895 M * Bertl well, I do not think that there are 'optimal' solutions in general 1162260913 M * Bertl you ahve all the 'options', you have to figure whats best for you :) 1162260973 M * Piet that's right, and I know my goals, but i'm not sure which way i need to go to reach my goals :) 1162261014 M * Piet i thinki'll just set up /var/lib/vservers on a seperate partition and allocate as much as possible space to it. 1162261023 M * Piet thanks for your help so far. 1162261205 M * Bertl np 1162261915 J * _mcp ~hightower@wolk-project.de 1162261915 Q * mcp Read error: Connection reset by peer 1162262278 M * Piet Bertl: do you think this is a bad idea for the partitioning of the host system for some reason? http://pastebin.co.uk/4879 1162262425 M * Piet i hope i'm not abusing your time 1162262425 M * Bertl if that's enough for you, why not ... 1162262434 M * Piet ok 1162262445 M * Bertl most distros will use up more space in /usr though 1162262460 M * Piet its debian etch on a server 1162262472 M * doener yep, and probably a lot less in var (assuming that most action happens in the vserver) 1162262489 M * Bertl ah, well, var seems okay, after all it contains logs 1162262508 M * Piet i was thinkign about finding a way to have the vserver log to the host 1162262513 M * Bertl (and at least I tend to keep logs for a longer time :) 1162262518 M * doener hm, my /var uses 1.2G and that includes my 32bit chroot... 1162262533 M * Bertl Piet: that's quite easy, just enable network logging 1162262552 M * Piet using syslog? 1162262553 M * doener and I wonder what /lib/init/rw is, and why a tmpfs is mounted there... *goes searching* 1162262563 M * Piet i guess i'll need to look into this later 1162262598 M * Piet i totally forgot about /usr :) 4 GB should be fine on top of the 450 MB for / ? 1162262664 M * doener unless you want to compile in /usr/src 1-1.5GB are probably more than enough (using 995MB here incl. X, fvwm etc.) 1162262704 M * Piet i may need to compile a kernel but i think this doesn't take too much space 1162262713 M * doener you can compile that in /home ;) 1162262729 M * Piet right :) 1162262750 M * doener anyway, off to bed now... g'night! 1162262759 M * Piet and you 1162262857 M * Bertl night doener! 1162263376 Q * ensc Killed (NickServ (GHOST command used by ensc_)) 1162263386 J * ensc ~irc-ensc@p54B4D168.dip.t-dialin.net 1162264054 Q * bronson Ping timeout: 480 seconds 1162264209 J * s0undt3ch_ ~s0undt3ch@81.193.57.87 1162264640 Q * s0undt3ch Ping timeout: 480 seconds 1162264640 N * s0undt3ch_ s0undt3ch 1162264682 J * FireEgl FireEgl@Sebastian.Atlantica.US 1162266239 J * s0undt3ch_ ~s0undt3ch@81.193.57.175 1162266670 Q * s0undt3ch Ping timeout: 480 seconds 1162266670 N * s0undt3ch_ s0undt3ch 1162268153 M * Bertl okay, off to bed now .. have fun everyone! 1162268164 N * Bertl Bertl_zZ 1162268367 Q * ms_ Ping timeout: 480 seconds 1162268565 Q * sladen Ping timeout: 480 seconds 1162268613 J * sladen paul@starsky.19inch.net 1162268997 J * s0undt3ch_ ~s0undt3ch@81.193.56.228 1162269112 Q * s0undt3ch Ping timeout: 480 seconds 1162269112 N * s0undt3ch_ s0undt3ch 1162275096 J * ms_ ~ms@arkansas.doc.ic.ac.uk 1162276892 Q * Aiken Ping timeout: 480 seconds 1162276989 J * s0undt3ch_ ~s0undt3ch@81.193.58.70 1162277330 Q * s0undt3ch Read error: Operation timed out 1162277330 N * s0undt3ch_ s0undt3ch 1162277838 J * Piet_ hiddenserv@tor.noreply.org 1162278191 Q * Piet Ping timeout: 480 seconds 1162278590 J * bronson ~bronson@c-71-198-75-160.hsd1.ca.comcast.net 1162278967 J * meandtheshell ~markus@85-124-232-84.work.xdsl-line.inode.at 1162279730 J * Zaki_ ~barata@212.118.106.224 1162280115 Q * Zaki Ping timeout: 480 seconds 1162281941 J * dna_ ~naucki@0-222-dsl.kielnet.net 1162283787 M * renihs morning 1162283842 Q * ms_ Ping timeout: 480 seconds 1162284135 J * rgl Rui@217.129.151.190 1162284139 M * rgl good morning 1162284990 J * Aiken ~james@tooax6-249.dialup.optusnet.com.au 1162285380 J * Zaki[] ~barata@212.118.110.86 1162285758 Q * shedi Quit: Leaving 1162285760 Q * Zaki_ Ping timeout: 480 seconds 1162286003 Q * michal` Ping timeout: 480 seconds 1162286973 M * rgl how do I pass an argument to the guest init? 1162287084 J * michal` ~michal@www.rsbac.org 1162287161 M * rgl hi michal` 1162287217 Q * eGnarF Read error: Operation timed out 1162287358 J * eGnarF ~bartek@bk.crystone.se 1162288458 J * debugger ~Rui@217.129.151.190 1162288907 Q * rgl Ping timeout: 480 seconds 1162289495 J * cunha ~adray@109.102.broadband6.iol.cz 1162289923 M * cunha Hi, how can I test cpu scheduling? I've set (1/4 of cpu for the first a 3/4 for the second vserver) and files and I have two identical vservers. I tried to run two identical programs written in C, each in one vserver. It does some float computations. Every 1000 cycles it prints dot. But both instances print the same amount of dots in the same time. Any idea, please? 1162290329 J * GoNoGo ~GoNoGo@cro34-3-82-236-93-215.fbx.proxad.net 1162290571 J * ms_ ~ms@arkansas.doc.ic.ac.uk 1162291078 Q * virtuoso Read error: Operation timed out 1162291100 J * virtuoso ~s0t0na@shisha.spb.ru 1162292182 J * dlezcano ~dlezcano@cimai.net4.nerim.net 1162292188 M * dlezcano hi all 1162292826 N * _mcp mcp 1162292990 N * mcp _mcp 1162293597 Q * ms_ Ping timeout: 480 seconds 1162293605 Q * Aiken Quit: Leaving 1162293999 Q * Skram Ping timeout: 480 seconds 1162294067 N * debugger rgl 1162294070 M * rgl yellow 1162294222 Q * hardwire Ping timeout: 480 seconds 1162295608 M * derjohn cunha, there is a tool cpuhog.c , which computes "power of" and consumes 100% CPU 1162295648 J * debugger Rui@217.129.151.190 1162295725 Q * debugger 1162296117 Q * rgl Ping timeout: 480 seconds 1162296591 M * renihs or burncpu 1162296595 M * renihs cpuburn 1162297325 M * derjohn cpuhog is what Bertl coded :) 1162297332 M * renihs oh 1162297335 M * renihs than cpuhog it is 1162297356 M * renihs why isnt that thingie not in gentoo portage? :) 1162297382 M * derjohn maybe to prevent abuse ? 1162297432 M * renihs ? 1162297435 M * renihs how do you mean? 1162297441 M * derjohn http://vserver.13thfloor.at/Experimental/TOOLS/cpuhog.c 1162297465 M * derjohn well, If you start it on a host, e.g. by accident, the cpu will be at 100% all the time 1162297494 M * derjohn but no particular exploit or such ... maybe silly users start it in a unlimited guest 1162297585 M * renihs ya but the same would be accomplished if i calculate pi or use cpuburn? 1162297600 M * derjohn indeed, yes 1162297623 M * renihs :) 1162298120 M * Hollow no 2.1.1? :( 1162298250 M * renihs 2.1.1? 1162298277 M * Hollow yeah.. two days ago it was said to be released during the night.. ;) 1162298283 M * Hollow but well :) 1162298350 M * derjohn Hollow, there might be issues: doener and Bertl_zZ know more. But probably rc45 wont be the release. rc46 ? I dunno. 1162298412 M * Hollow i see... 1162298471 M * harry what's wrong with a cpuhog like this: int main(){asm("jmp -1");} 1162298707 J * bauer ~bauer.fre@202.120.43.244 1162299125 M * doener derjohn: I'm not sure if the issues sla saw are vserver related 1162299222 J * Skram ~Mark@hermes.sentiensystems.com 1162300326 J * ms_ ~ms@arkansas.doc.ic.ac.uk 1162300666 Q * Hunger Ping timeout: 480 seconds 1162300891 A * phreak`` pokes _rc45 1162300895 M * phreak`` won't even boot ... 1162300953 Q * eyck_ Remote host closed the connection 1162301114 Q * pagano Ping timeout: 480 seconds 1162301308 J * Hunger Hunger.hu@213.163.11.138 1162301312 P * bauer 1162301433 J * eyck eyck@ghost.anime.pl 1162301545 Q * Piet_ Quit: Piet_ 1162302775 M * derjohn phreak``, my rc45 boots ... do you refer to the hosts boot process or the guests? 1162303578 M * blizz after unification, are libraries loaded only once into memory when they're the same? 1162303849 M * doener blizz: yes 1162303899 M * doener it's one file, just available under different names, so the kernel won't load it twice 1162304747 M * Hollow daniel_hozac: around? 1162304756 M * daniel_hozac heh, just got back. 1162304768 M * Hollow :) 1162304795 M * Hollow there is a bug in the $no_dev_sanity check.. and it made me think if the dev sanity would make sense for all build methods? 1162304817 M * daniel_hozac oh? 1162304830 M * Hollow yeah... it should be && instead of || 1162304841 M * Hollow broken util-vserver test logic :P 1162304848 M * Hollow confused me when making the patch 1162304872 M * daniel_hozac ah, yes. 1162304924 M * Hollow anyway, would it make sense to apply it to all build methods? how does yum/apt/etc handle device files in their packages? 1162304996 M * daniel_hozac debootstrap moves the util-vserver generated /dev to /dev.X and then replaces the debootstrap generated one, i think. 1162305020 M * daniel_hozac yum/apt/etc. i'm not sure. 1162305041 M * daniel_hozac i don't think they do anything about that. 1162305103 M * Hollow ok, so it doesn't really make sense for those.. 1162305108 M * Hollow will fix the template method then 1162305927 J * Ben81 ~Ben81@129.175.7.177 1162307063 J * mejo ~jonas@dslb-084-058-125-037.pools.arcor-ip.net 1162307068 M * mejo hello, 1162307088 M * mejo i'm running debian/unstable with a selfcompiled linux-image-2.6.18-4-amd64-resivo 1162307120 M * mejo now i would like to try out vserver, but i cannot find the kernel options in 'make menuconfig'. 1162307149 M * daniel_hozac have you applied the patch? 1162307199 M * mejo according to http://svn.debian.org/wsvn/kernel/dists/trunk/linux-2.6/debian/patches/, vs2.0.2.2-rc2.patch is applied to linux-source-2.6.18 1162307223 M * mejo daniel_hozac: the debian package linux-source-2.6.18 should already have the vserver patch applied. 1162307245 M * mejo can you give me an example file in the linux source tree, which exists with the patch applied, and does not exist with the patch unapplied? 1162307246 M * daniel_hozac the Debian kernel has lots of flavors. you need to specify which one you want 1162307260 M * daniel_hozac the kernel/vserver, for instance. 1162307267 M * mejo how do i do so? 1162307282 M * daniel_hozac that i don't know, i'm a Fedora guy. waldi, micah? 1162307289 M * daniel_hozac or derjohn? 1162307335 M * mejo but vserver is not an own architecture (like xen is), is it? 1162307354 M * daniel_hozac no. 1162307384 M * mejo for xen i need to use 'make-kpkg --arch xen', IIRC. 1162307423 M * mejo but how do i select a flavor? i cannot find anything about 'flavour' or 'vserver' in the make-kpkg manpage. 1162307442 M * daniel_hozac as i said, i have no idea. 1162307639 M * brcc w 1162307651 M * brcc without limiting VM(AS), is it possible for a vserver to crash the whole machine ? 1162307694 Q * nox Quit: Lost terminal 1162307697 J * nox ~nox@static.88-198-17-175.clients.your-server.de 1162307714 M * daniel_hozac AS is not the most relevant limit there... 1162307811 M * daniel_hozac and what does "crash" mean in this instance? 1162308015 M * brcc swapping so much that stopped responding 1162308039 M * brcc i was monitoring with vmstat, it usually does fre swapping 1162308041 M * brcc suddenly toom uch swap 1162308085 M * brcc and stopped responding.. (ping was still working) 1162308088 M * brcc i think it was trashing 1162308117 M * brcc So i was wondering if it was possible that some vserver did that 1162308319 Q * Ben81 Quit: Leaving 1162308795 M * eGnarF brcc: it is. I did it myself by mistake once. 1162308948 M * brcc How can we fix it ? limiting as ? 1162308976 M * eGnarF brcc: That's what I did. 1162308994 M * eGnarF brcc: seems to work. 1162309015 M * phreak`` derjohn: host, it barfs about / being mount ro 1162309202 M * ensc Hollow: /dev is added to rpm's %_netsharedpath which prevents that packages place files into /dev; current distros are udev based and the /dev issue should not matter anymore 1162309274 M * Hollow ensc: ok, thanks! 1162309275 M * cunha How can I set the number of CPUs for vserver? Thx 1162309291 M * daniel_hozac cunha: cpusets. 1162309435 M * derjohn phreak``, interesting one of my amd64 host has similar mount issues 1162309457 M * derjohn phreak``, but my p3-notebook setup runs fine 1162309462 M * cunha daniel_hozac: Well, I haven't found what to write in. 1162309505 M * daniel_hozac derjohn: new since -rc45? or since 2.6.18? 1162309522 M * daniel_hozac (because i had problems booting my 2.6.18 kernel on the x86_64 server as well) 1162309615 M * daniel_hozac cunha: [-] in cpus, and [-] in mems, and whatever you want in name. 1162309648 M * daniel_hozac cunha: that assumes you have cpuset support already though, you might have to mkdir -p /dev/cpuset; mount -t cpuset none /dev/cpuset 1162309838 N * Bertl_zZ Bertl 1162309845 M * Bertl morning folks! 1162309848 M * daniel_hozac ensc: btw, could you push revision 2356 of util-vserver to FC-6 and devel? (maybe with a patch?) would be nice not having to patch manually to use a fixed yum. 1162309852 M * daniel_hozac morning Bertl! 1162310088 M * brcc eGnarF: thanks egnarf! 1162310100 M * brcc eGnarF: how big did you limit it ? 1162310237 M * eGnarF brcc: 512mb was enough for my needs. But remember that the limits are in number of pages, not in megabytes. 1162310237 Q * FireEgl Read error: Connection reset by peer 1162310261 M * brcc sure. i just want to make sure no one will crash the machine 1162310262 P * cunha 1162310277 M * brcc Is there a way to force swapping so i can test this limit ? 1162310463 M * Bertl there is soemthing called swappiness (kernel parameter) 1162310480 M * Bertl adjusting this one will make the system swap out more/less agressive 1162310543 M * brcc great. gonna try that out. thanks a lot ebrtl! 1162310886 M * phreak`` derjohn: I'm going to take a look what changed between -rc44 and -rc45 ext3 wise 1162311001 M * phreak`` (or ask Bertl) 1162311015 M * daniel_hozac phreak``: nothing. 1162311035 M * phreak`` daniel_hozac: then something other in that regard has changed :) 1162311052 M * daniel_hozac so it worked fine in -rc44? 1162311105 M * phreak`` daniel_hozac: it works fine when I boot -rc44 (thats what currently is running); when I try to boot -rc45 it somehow assumes it should mount / ro, thus the whole Gentoo init fails :) 1162311123 M * renihs -rc44? 1162311133 M * daniel_hozac different kernel options? 1162311133 M * renihs dont you dare mistreat your gentoo :) 1162311137 M * Bertl hum? sure you didn't add 'ro' instead of 'rw' on the cmdline? 1162311151 M * phreak`` $ cat /boot/grub/grub.conf | grep kernel 1162311151 M * phreak`` kernel (hd0,0)/boot/kernel-2.6.18.1-vserver-2.1.1-rc45 root=/dev/hda1 quiet vga=0x31A 1162311154 M * phreak`` kernel (hd0,0)/boot/kernel-2.6.18.1-vserver-2.1.1-rc44 root=/dev/hda1 quiet vga=0x31A 1162311165 M * daniel_hozac initrd? 1162311168 M * phreak`` nope 1162311175 M * daniel_hozac for neither? 1162311183 M * phreak`` yeah, for none of my systems 1162311186 M * phreak`` :) 1162311190 J * hardwire ~hardwire@89-208-58-66.gci.net 1162311203 M * Bertl hmm, that should result in an ro root indeed, IIRC 1162311219 M * Bertl userspace has to remount it later 1162311283 A * phreak`` goes downstairs to C&P the output from the terminal 1162311315 M * phreak`` iirc, it tries to remount it rw, but somehow fails. 1162311324 M * phreak`` brb 1162311507 J * lilalinux ~plasma@dslb-084-058-220-007.pools.arcor-ip.net 1162311752 Q * eyck Ping timeout: 480 seconds 1162312559 M * phreak`` Bertl: it really looks like on -rc44 / get remounted by gentoo's baselayout; on -rc45 not .. 1162312565 Q * dna_ Read error: Connection reset by peer 1162312583 J * dna_ ~naucki@89.27.222.0 1162312682 M * Bertl phreak``: let's specify rw on boot and check for differences, if possible (i.e. differences in how it booted up/what was checked and so on) 1162312693 M * Hollow apropos gentoo baselayout... this thing drives me nuts... 1162312700 M * Hollow it gets worse every day 1162312713 M * phreak`` lol, tried to get 1.13 working in a vps ? 1162312724 M * Hollow yep... 1162312733 M * Hollow it is kind-of working 1162312739 M * Hollow but this fifo thing just drives me nuts 1162312762 M * phreak`` Hollow: lol, ok; same stage where I was before I was ressurecting 1162312889 M * phreak`` Bertl: same result as before .. 1162312900 M * phreak`` http://paste.linux-vserver.org/598 <-- -rc44 1162313098 J * eyck_ eyck@ghost.anime.pl 1162313336 M * Bertl phreak``: and for rc45? 1162313459 M * phreak`` Bertl: sec, I'm still typing it from the screen ... 1162313478 M * phreak`` (no serial console available ...) 1162313618 M * phreak`` Bertl: http://paste.linux-vserver.org/599 1162313797 M * Bertl ah, okaywhat's in line 103 and 109? 1162313824 M * Hollow phreak``: jabber? 1162313890 M * phreak`` Hollow: in a sec :) 1162314165 M * phreak`` Bertl: according to vim, http://paste.linux-vserver.org/600 1162314205 M * Hollow haha... else 1162314215 M * Hollow serious killer on a ro filesystem 1162314399 J * stefani ~stefani@tsipoor.banerian.org 1162314628 M * Bertl phreak``: hum? what does that mean? 1162314643 Q * bronson Ping timeout: 480 seconds 1162314652 J * bonbons ~bonbons@83.222.36.111 1162314719 M * phreak`` Bertl: hrm ? the meaning of my paste or the meaning of my sentence before the paste ? :) 1162314725 A * phreak`` 's a bit confused :P 1162314736 M * Hollow it does mean that your system is really screwed 1162314755 M * Hollow ;) 1162314756 M * phreak`` Hollow: with -rc45, yeah. I *perfectly* works with -rc44 :P 1162314768 M * phreak`` s/. I/. it/ 1162314775 M * Hollow well, probably i should try rc45... 1162314785 M * phreak`` you better do :P 1162314809 A * Hollow goes compiling 1162314819 M * Bertl well, the only relevant changes are the cow fixes from daniel 1162314827 M * Hollow restart is good anyways... i have a dozen defunct inits around :) 1162314845 M * Bertl everything else should not affect you 1162314937 M * Hollow daniel_hozac: btw, do you see any chance that we can give multiple commands to INITCMD_START somehow? 1162314968 M * daniel_hozac Hollow: not really, i guess a baselayout wrapper script would do though, right? 1162314978 M * Hollow daniel_hozac: probably, yeah.. 1162314984 M * Hollow will have to try 1162315012 M * Hollow will commit some fixes for the gentoo INIT_POST anyways... 1162315091 J * s0undt3ch_ ~s0undt3ch@bl9-225-108.dsl.telepac.pt 1162315515 Q * s0undt3ch Ping timeout: 480 seconds 1162315515 N * s0undt3ch_ s0undt3ch 1162315782 Q * ms_ Ping timeout: 480 seconds 1162316342 Q * eyck_ Ping timeout: 480 seconds 1162316358 M * phreak`` Bertl: I think you need to take a look at the changes to fs/open.c, because that is your/our bugger :) 1162316379 M * daniel_hozac phreak``: what's wrong with it? 1162316419 M * daniel_hozac oooh, you don't have cow link breaking? 1162316419 M * phreak`` it somehow makes my system "misbehave" .. 1162316427 M * phreak`` nope, I dont have :) 1162316428 M * daniel_hozac shit, i totally forgot about that case. 1162316433 M * phreak`` *g* 1162316441 M * daniel_hozac sorry about that. 1162316443 M * phreak`` thanks a lot then :) 1162316457 M * phreak`` daniel_hozac: no problem at all ;) 1162316470 M * Bertl hehe, good that we didn't release :) 1162316476 M * daniel_hozac indeed. 1162316494 M * phreak`` Bertl: btw, the pid_task oops is still there :) (guess that wasn't addressed in -rc45). 1162316509 M * Bertl no oops, just a warning 1162316519 M * Bertl (i.e. BUG: warning vs BUG at 1162316536 M * phreak`` yeah, BUG :) 1162316544 M * daniel_hozac the trace is supposed to be removed before release. 1162316603 M * phreak`` ok, good to know :) 1162316604 M * daniel_hozac is gcc clever enough to optimize the function properly so that we can do the #ifdef's inside the function instead? 1162316615 M * daniel_hozac and get the same result, i mean. 1162316621 J * Rusty` ~rusty@a1052.adsl.pool.eol.hu 1162316624 M * Rusty` hello 1162316626 M * daniel_hozac hi 1162316636 M * Rusty` i have a little problem :) 1162316649 M * daniel_hozac tell us all about it. 1162316656 M * Rusty` :) 1162316656 M * Rusty` apollo:/vservers# vserver apache start 1162316657 M * Rusty` Cannot find device "eth0:1" 1162316657 M * Rusty` CONSOLE_LOG_LEVEL is of unaccepted value. 1162316657 M * Rusty` KERNEL_RINGBUF_SIZE is of unaccepted value. 1162316691 M * Rusty` apollo:~# ifconfig |grep eth0:1 1162316691 M * Rusty` eth0:1 Link encap:Ethernet HWaddr 00:50:8B:4B:45:16 1162316691 M * Rusty` apollo:~# 1162316713 M * daniel_hozac eth0:1 isn't a valid device. 1162316721 M * daniel_hozac you should put eth0 as device, and 1 as the name. 1162316738 M * daniel_hozac and i assume you meant to use nodev as well, right? 1162316773 M * Rusty` i have two ip address 1162316783 M * Rusty` iface eth0 inet static 1162316784 M * Rusty` address 10.1.1.200 1162316784 M * Rusty` netmask 255.255.255.0 1162316784 M * Rusty` iface eth0:1 inet static 1162316784 M * Rusty` address 195.38.113.36 1162316785 M * Rusty` netmask 255.255.255.192 1162316785 M * Rusty` gateway 195.38.113.62 1162316821 M * Rusty` i would like to use the "public" ip in the vserver 1162316840 M * Rusty` and the "privat" ip in mother server 1162316868 M * daniel_hozac so put 195.38.113.36 in ip, 26 in prefix, and touch nodev. 1162316884 M * Rusty` ah so 1162316924 M * daniel_hozac you've already setup the address on the host, dev is just if you want the utils to create/remove the address. 1162316971 M * Rusty` it works, thanks :) 1162316972 M * Rusty` CONSOLE_LOG_LEVEL is of unaccepted value. 1162316972 M * Rusty` KERNEL_RINGBUF_SIZE is of unaccepted value. 1162316979 M * Rusty` is this a big problem? 1162316996 M * Bertl just means that you haven't enabled syslog virtualization :) 1162317008 M * Bertl (and you guest has scripts runnign to mess with that) 1162317013 M * phreak`` Bertl: lol :) 1162317061 M * daniel_hozac phreak``: http://people.linux-vserver.org/~dhozac/p/k/delta-cow-fix07.diff ? 1162317107 M * Rusty` syslog virtualization? what is this? 1162317107 M * phreak`` daniel_hozac: that on top of -rc45 ? 1162317113 M * daniel_hozac phreak``: yep. 1162317139 M * phreak`` Rusty`: he's joking. you need to get rid of those default init-scripts, that assume you have hardware available inside the vps 1162317154 M * Bertl http://linux-vserver.org/Capabilities_and_Flags 1162317162 M * Bertl nah, you can really enable this ... 1162317202 M * Bertl hmm, seems it's not lsited there, sec 1162317216 M * Hollow ahem 1162317218 M * Rusty` o:) 1162317225 M * daniel_hozac Bertl: does the patch look sane to you? 1162317231 M * derjohn What? Syslog v12n ? An own dmeg per guest ?? 1162317235 M * Hollow Bertl: 12 0x00001000 SYSLOG Allow syslog(2) 1162317245 M * daniel_hozac derjohn: yes, but nothing logs there. 1162317273 M * Rusty` how can i enable this? :) 1162317294 M * Hollow echo syslog >> /etc/vservers//ccapabilities 1162317297 M * derjohn ah, a syslog "dummy". well, I thought thingies like "resource limit hit" might be listed there 1162317316 M * Bertl Hollow: ah, tx, didn't see it :) 1162317334 M * Hollow :) 1162317350 M * Bertl daniel_hozac: have to leave now, will double check later 1162317351 M * Rusty` apollo:~# vserver apache start 1162317351 M * Rusty` Unknown ccap 'syslog' 1162317359 M * Hollow uhm... util-vserver madness 1162317362 M * Hollow Rusty`: sec.. 1162317364 M * daniel_hozac Bertl: ok, np. 1162317371 M * Bertl cya folks! 1162317377 N * Bertl Bertl_oO 1162317377 M * Rusty` it is an old util-vserver 1162317403 M * daniel_hozac Rusty`: that'd be why then. 1162317431 M * Hollow ah ok.. and i thought it has a different name... ;) 1162317457 M * daniel_hozac got me curious too, i thought we had already found all the ones with strange names ;) 1162317471 M * Hollow yeah, i hope we did :) 1162317528 M * Hollow daniel_hozac: though util-vserver is based on 2.0, it might not hurt to put in the newer ccaps and cflags, right? so i don't have to lookup numbers all the time... ^38 sucks :) 1162317538 M * daniel_hozac :P 1162317555 M * daniel_hozac ^38 is persistent though, no? 1162317558 M * Hollow yep.. 1162317563 M * Hollow just an example.. 1162317564 M * daniel_hozac that should be in 0.30.211, and in stable... 1162317586 M * Hollow ah i see.. 1162317590 M * Hollow bad example 1162317591 M * Hollow :) 1162317593 M * daniel_hozac :) 1162317607 M * Hollow but good to know 1162317646 M * daniel_hozac ADMIN_MAPPER, ADMIN_CLOOP, VIRT_TIME, are the only devel specific ones, no? 1162317741 M * Hollow seems so yes... INFO_SCHED is missing too, but well, legacy doesn't count :P 1162317777 M * daniel_hozac INFO_LOCK should probably be updated... 1162317802 M * daniel_hozac or maybe i'm misunderstanding. 1162317822 M * Hollow isn't it somewhat the same as ~STATE_ADMIN? 1162317823 M * daniel_hozac does the current description mean it's impossible to migrate to it? or that _it_ can't migrate? 1162317841 M * Hollow if the description is right it means you cannot migrate to it 1162317857 M * daniel_hozac well, ~STATE_ADMIN is a lot more than just migrating. 1162317867 M * Hollow right... 1162317874 M * Hollow but why is it legacy then? 1162317875 M * daniel_hozac ok, so that's correct then, and i just misunderstood the description :) 1162317882 M * daniel_hozac it's not anymore :) 1162317892 M * Hollow eiei 1162317895 M * daniel_hozac it is a legacy flag though, like it has been around forever. 1162317898 A * Hollow goes updating the wiki 1162317901 M * daniel_hozac (AFAIK) 1162317911 M * Hollow i.e X instead of L? 1162317928 M * daniel_hozac hmm, well, the meaning is the same, isn't it? 1162317928 M * Hollow or it has just been re-implemented? 1162317932 M * Rusty` ah it works, thanks 1162317933 M * Rusty` :) 1162317968 M * Hollow daniel_hozac: indeed... /me confused 1162317972 Q * dlezcano Quit: Leaving 1162317977 M * daniel_hozac me too. 1162317980 M * Hollow :) 1162317986 M * phreak`` daniel_hozac: yeah, that on top of -rc45 works for me :) 1162317989 M * daniel_hozac i guess it's fine as is :) 1162317995 M * daniel_hozac phreak``: ok, thanks. 1162317995 M * Hollow probably.. 1162318003 M * phreak`` daniel_hozac: thank you ! 1162318024 M * Hollow looks like today is bug fixing day 1162318026 M * Hollow ;) 1162318031 M * phreak`` hah ;) 1162318034 M * Hollow wrt baselayout *ahem* 1162318041 M * phreak`` yeah, figured that :) 1162318047 M * Hollow but hey... sysinit and boot runlevel already work 1162318056 M * Hollow default still sucks.. 1162318056 M * phreak`` good then *g* 1162318092 M * Hollow we will probably add the syslog stub to gentoo initpost too... since baselayout messes with dmesg -n1 1162318113 M * Hollow and we will reimplement the gentoo init style 1162318136 M * Hollow together with some other initpost madness ;) 1162318163 J * bronson ~bronson@66.160.177.216 1162318193 M * phreak`` *ugh* that sounds like a awful lot of work :P 1162318217 M * Hollow nah.. pretty straight.. 1162318240 M * Hollow one of the next ebuilds will also deprecate vserver-new 1162318252 M * Hollow superceeded by vserver ... build -m template -d gentoo 1162318261 M * daniel_hozac +-- 1162318273 M * Hollow +--? 1162318291 M * daniel_hozac -m template -- -d gentoo 1162318298 M * Hollow <- lazy 1162318298 M * Hollow ;) 1162318320 M * daniel_hozac yeah, but it's required ;) 1162318328 M * Hollow i know... already figured that out today :P 1162318668 J * eyck eyck@ghost.anime.pl 1162318749 M * Hollow daniel_hozac: any reason why /proc/modules is in vprocunhide-files? 1162318786 M * Hollow also kcore does not seem too good..? 1162318796 M * daniel_hozac what? kcore is in there? 1162318806 M * Hollow nod. 1162318824 M * daniel_hozac ... ensc? 1162318862 M * Hollow some others i'm not really sure, if they are good to have... pci, iomem, ksyms? 1162318883 M * daniel_hozac yeah, seems rather strange. 1162318906 M * ensc kcore should be protected by the kernel 1162318907 M * ensc cat /proc/kcore 1162318907 M * ensc cat: /proc/kcore: Operation not permitted 1162318923 M * daniel_hozac but then what's the point of showing it inside the guest? 1162318926 M * ensc dunno, why modules should not be visible 1162318956 M * daniel_hozac security through obscurity? :) 1162318969 M * Hollow heh. 1162318983 M * daniel_hozac how did you assemble the current list? empirical testing of what various programs look for? 1162318987 M * Hollow in this special case, gentoos baselayout tries to load modules based on the existance of /proc/modules 1162319007 M * Hollow sure, EPERM all the time, but ugly startup messages 1162319024 M * ensc I removed the entries which were known for problems (e.g. sysrq) 1162319038 M * ensc all other were kept 1162319044 M * daniel_hozac ok. 1162319091 M * ensc I wanted to provide an environment like on a regular machine 1162319109 M * Hollow yeah, sounds reasonable 1162319112 M * ensc (perhaps some program checks for /proc/kcore to decide whether /proc is mounted) 1162319184 M * phreak`` hrm, shouldn't /proc/sys/vserver be in that file ? :) 1162319200 M * phreak`` or is that needed from *within* a vserver ? 1162319230 M * Hollow i guess you cannot change it from inside 1162319286 J * adrien ~mirc@ppp08-89.dsl.citenet.net 1162319292 M * adrien hi 1162319309 M * daniel_hozac indeed, EPERM. 1162319329 M * Hollow applies to all sysctl things IIRC 1162319386 M * phreak`` but somehow hiding the dirs specified in vprocunhide-files doesn't work .. (/proc/sys/debug and its content is still visible) 1162319482 M * Hollow dinner time... back later 1162319592 M * daniel_hozac phreak``: hmm, works fine here. 1162320456 Q * Osgiliath Quit: The Sign Of Wrath Awaked. 1162320807 J * Osgiliath ~osgiliath@kurzweil.no-ip.org 1162321902 Q * sladen Ping timeout: 480 seconds 1162322150 J * sladen paul@starsky.19inch.net 1162322511 J * ms_ ~ms@arkansas.doc.ic.ac.uk 1162322535 J * Piet hiddenserv@tor.noreply.org 1162323637 P * stefani I'm Parting (the water) 1162324221 M * Rusty` apollo:/etc/init.d# vserver apache start 1162324221 M * Rusty` procfs-security. Please read the FAQ for more details 1162324221 M * Rusty` http://www.linux-vserver.org/index.php?page=Linux-Vserver+FAQ 1162324222 M * Rusty` :( 1162324384 M * brcc Rusty`: /etc/init.d/vprocunhide start 1162324456 M * Rusty` oh, i have not /etc/init.d/vprocunhide 1162324473 M * hardwire I have not it either 1162324490 M * Rusty` after all 1162324518 M * doener Rusty`: debian? 1162324561 M * Rusty` yes 1162324615 M * daniel_hozac /etc/init.d/vserver start then ;) 1162324671 M * Rusty` aaaaaaaaaaaaaaaaa 1162324697 M * Rusty` vservers-default init script restarted the server ?! 1162324702 M * Rusty` :) 1162324704 J * shedi ~siggi@130.208.221.254 1162324722 M * daniel_hozac hmm, it has a vservers-default script but no vprocunhide script? 1162325908 Q * Piet Remote host closed the connection 1162325941 J * Piet hiddenserv@tor.noreply.org 1162327141 Q * lilalinux Remote host closed the connection 1162328714 M * Hollow it seems VXC_SECURE_MOUNT does not enable the ability to mount tmpfs :/ 1162329205 M * daniel_hozac hmm. works with binary_mount too though, right? 1162329217 M * Hollow ah.. have to try.. 1162329240 M * Hollow yep.. works 1162329442 J * cunha ~adray@109.102.broadband6.iol.cz 1162329574 Q * GoNoGo Quit: beam me up sco... 1162329585 M * cunha Hi, I'm triing to use cpu-scheduling. It doesn't work. I wrote necessary lines into and files. What's wrong? 1162329670 M * daniel_hozac are you using hard CPU scheduling? 1162329679 M * daniel_hozac and in what way doesn't it work? 1162329761 M * Hollow phreak``: http://paste.linux-vserver.org/602 :)) 1162329778 M * daniel_hozac it works? :) 1162329781 M * Hollow indeed 1162329788 M * daniel_hozac wee. 1162329826 M * cunha Yes. The two identical programs in two identical vservers (the first with 1/4 of cpu, the second wtih 3/4 of cpu) run same speed. 1162329830 M * Hollow many things work here: 1) gentoo init style, 2) native baselayout, i.e. no baselayout-vserver anymore, 3) baselayout actually mounting stuff... 1162329847 M * daniel_hozac cunha: so sched_hard in the flags file, and the feature enabled in the kernel config? 1162329858 M * daniel_hozac Hollow: wow, awesome! 1162329892 M * daniel_hozac so we'll resurrect Gentoo initstyle for 0.30.212 after killing it in 211? :) 1162329897 M * Hollow this will obsolete so much workaround stuff in our tree... yay! 1162329909 M * Hollow daniel_hozac: heh, yeah.. 1162329913 M * cunha daniel_hozac: "...the feature enabled in the kernel config" - probably that's it. I haven't found it. Thanx 1162329924 M * Hollow daniel_hozac: but... 1.13.0 will probably still take quite some time to get stable 1162329940 M * daniel_hozac Hollow: ah, and it works only with 1.13.0? 1162329970 M * Hollow yep... if i'm bored i probably backport it to 1.12, but not sure it will even be accepted "upstream" 1162330007 M * daniel_hozac doesn't 1.12 require baselayout-vserver? aren't you upstream? 1162330037 M * Hollow well, i'm upstream of baselayout-vserver, but not of baselayout... i have no commit access etc.. 1162330057 M * daniel_hozac yeah, but does baselayout 1.12 work in guests at all? 1162330062 M * Hollow and the really great thing here is, that the patches to baselayout-1.13 will obsolete baselayout-vserver 1162330077 M * Hollow yep, it does work, although there seem to be some bugs with fifos.. 1162330087 M * Hollow but in general it works 1162330094 M * daniel_hozac ah, the strange fifo bug. 1162330105 M * daniel_hozac have you been able to track that down yet? 1162330151 M * Hollow not quite, though i got an idea while fixing 1.13... the issue seems to be caused by cruft in /var/lib/init.d not cleaned out during the sysinit runlevel 1162330159 M * Hollow will nivestigate that tomorrow probably 1162330179 M * daniel_hozac sounds good. 1162330342 M * Hollow daniel_hozac: the only thing here is: do you think it is ok to force secure_mount, binary_mount on gentoo users? i mean.. it should be safe, right? 1162330389 M * Hollow although i think it should even work without with some minor improvements 1162330392 M * daniel_hozac well, secure_mount/binary_mount will let the guest mount anything. 1162330402 M * Hollow s/without// 1162330438 M * daniel_hozac binary_mount might lead to processes stuck in D state. 1162330453 M * daniel_hozac (NFS where the filer goes away) 1162330461 M * Hollow have to take special care for ovz here too, since they mount proc and devpts from inside... but we don't do that.. have to get the logic right here 1162330493 M * Hollow yeah.. especially binary_mount is a critical one IMO 1162330503 M * daniel_hozac hmm, proc and devpts should work with secure_mount, i think. 1162330512 J * Aiken ~james@tooax8-217.dialup.optusnet.com.au 1162330516 M * daniel_hozac i'm not sure if binary_mount is really the correct requirement for tmpfs. 1162330549 M * Hollow yeah.. tmpfs isn't really required though... i just saw /dev/shm in the default fstab, but we overwrite that in initpost anyway... 1162330636 M * Hollow the prolem i got with devpts is, that when trying to vlogin it says: openpty: operation not permitted 1162330644 M * Hollow when mounting from outside it works 1162330662 M * daniel_hozac ah, right, the forced MS_NODEV. 1162330678 M * Hollow yup.. 1162330764 M * Hollow anyway... working baselayout is f*cking awesome.. as i said.. it will deprecate so many things.. :D 1162330799 M * Hollow baselayout-vserver, the vserver portage profile, extra stages .. :) 1162330815 M * daniel_hozac hehe. 1162330905 M * Hollow not to forget the boot output... mayn folks seem to want that, and it's good for debugging 1162331506 Q * dna_ Quit: Verlassend 1162332072 P * cunha 1162335241 Q * meandtheshell Quit: exit (0); 1162336239 N * Bertl_oO Bertl 1162336243 M * Bertl evening folks! 1162336355 M * daniel_hozac evening Bertl! 1162336372 M * Bertl ah, so any feedback for your patch? 1162336427 M * daniel_hozac phreak`` said it fixed his problems, it seems to still work fine for me as well. 1162336463 M * Bertl okay, will check it in a few minutes 1162336613 M * Bertl daniel_hozac: okay, here some comments: 1162336644 M * Bertl - we should avoid duplicate assignments, is inode used on dput_adn_out? 1162336664 M * Bertl (if not, just move the inode = assignment down) 1162336705 M * Bertl - why does IS_COW(inode) trigger with CoW disabled? 1162336724 M * Bertl (IMHO that should evaluate to const 0 in that case, maybe a bug?) 1162336802 M * daniel_hozac so we'll move the #ifdef to include/linux/fs.h instead. 1162336809 M * daniel_hozac i don't get your first comment though. 1162336834 M * Bertl sec, let me double check with the code 1162336838 M * daniel_hozac oh, i see, in sys_fchmodat? 1162336845 M * daniel_hozac and do_utimes... 1162336856 M * Bertl yep 1162336857 M * daniel_hozac yeah, that first assingment could be removed. 1162336866 M * daniel_hozac now that it's unconditional. 1162336953 M * Bertl yep, looks like a plan 1162337265 M * daniel_hozac so http://people.linux-vserver.org/~dhozac/p/k/delta-cow-fix08.diff ? 1162337506 M * Bertl yep, looks good 1162337782 Q * ms_ Ping timeout: 480 seconds 1162338511 Q * Greek0 Quit: leaving 1162338547 J * Greek0 ~greek0@85.255.145.201 1162338849 J * FireEgl FireEgl@Sebastian.Atlantica.US 1162339060 M * hardwire delta cow 1162339067 M * hardwire I saw one of those in Tenessee 1162339124 M * Bertl and, did it work as expected? :) 1162339167 Q * Greek0 oxygen.oftc.net strange.oftc.net 1162339167 Q * bronson oxygen.oftc.net strange.oftc.net 1162339167 Q * Skram oxygen.oftc.net strange.oftc.net 1162339167 Q * Hollow oxygen.oftc.net strange.oftc.net 1162339167 Q * gdm oxygen.oftc.net strange.oftc.net 1162339167 Q * yang oxygen.oftc.net strange.oftc.net 1162339167 Q * Borg- oxygen.oftc.net strange.oftc.net 1162339167 Q * mugwump oxygen.oftc.net strange.oftc.net 1162339167 Q * duckx oxygen.oftc.net strange.oftc.net 1162339167 Q * anonc oxygen.oftc.net strange.oftc.net 1162339170 J * _ ~hollow@styx.xnull.de