1170374493 N * jbailey jbailey-afk 1170374550 J * dmax ~semaj@81.193.57.88 1170375119 N * Bertl_oO Bertl 1170375140 M * Bertl back now ... but I'm probably off to bed soon ... 1170375149 M * Bertl anything urgent or important? 1170375180 M * daniel_hozac nothing that i can think of. 1170375206 M * Bertl okay, good, any progress on the long atomic side? 1170375221 M * daniel_hozac no, i haven't gotten around to that yet, i'll do it tomorrow. 1170375230 M * Bertl np, whenever you feel like it ... 1170375534 M * Bertl okay, off to bed then .. have a good one everyone! 1170375538 M * daniel_hozac good night! 1170375541 N * Bertl Bertl_zZ 1170375544 M * Bertl_zZ you too! 1170376282 Q * dna Quit: Verlassend 1170377351 J * jjpacheco ~justin@CPE00146c1608af-CM00159a655a08.cpe.net.cable.rogers.com 1170377355 M * jjpacheco hola everyone! 1170377360 M * jjpacheco question 1170377408 M * jjpacheco is it possible to make a vserver re-mount its mounted partitions (described in its fstab) without restarting the vserver? 1170377433 M * jjpacheco also, i would like to keep the hosts mount table clean 1170377464 M * jjpacheco i tried vnamespace -e 130 mount -a 1170377469 M * jjpacheco but that just didn't work 1170377669 M * daniel_hozac vnamespace -e /usr/lib*/util-vserver/secure-mount .... 1170377973 M * jjpacheco that will force the vserver to remount its partitions? 1170378013 M * daniel_hozac what partitions are you looking to remount? 1170378045 M * daniel_hozac with the correct options, that should be possible though. 1170378062 M * daniel_hozac (and i think you need to make sure your cwd is the root of the guest when running that command) 1170378064 M * jjpacheco mount binds actually 1170378221 M * daniel_hozac i don't think you can remount bind mounts. 1170378225 M * Guy- daniel_hozac: is the wiki supposed to have a structure, or can I just go and create a new page that describes my vserver+runit setup? 1170378246 M * daniel_hozac if it's specific to util-vserver, it should be in the util-vserver: namespace. 1170378267 M * Guy- I'm not sure what qualifies as 'specific' 1170378286 M * Guy- it's a way of running services inside vservers supervised by a runit running on the host 1170378299 M * Guy- sure, util-vserver is involved 1170378343 M * daniel_hozac so just create a page then. 1170378364 M * Guy- OK 1170378384 M * Guy- can I reference your signal-relay.c at its current location? 1170378392 M * Guy- or are you going to include it in util-vserver? 1170378538 M * daniel_hozac i don't think it really belongs in util-vserver. 1170378591 M * Guy- OK, maybe I'll just include it in the page verbatim 1170378594 M * Guy- it's not long, after all 1170378644 M * daniel_hozac i'd prefer keeping it where it is. 1170378675 M * Guy- OK 1170378828 M * jjpacheco daniel_hozac: so i can't remount mount binds? 1170378844 M * jjpacheco i basically just want the vserver to re-read its fstab 1170378854 M * jjpacheco similar to mount -a (on the host system) 1170378893 M * daniel_hozac in what way? 1170378900 M * daniel_hozac do you want to add another mount? 1170378965 M * jjpacheco mmm 1170378979 M * jjpacheco well, i'd like to make an addition to its fstab 1170378992 M * daniel_hozac wouldn't it be easier to just do vnamespace -e mount...? 1170378994 M * jjpacheco then do somethign like mount -a 1170379010 M * jjpacheco it would, but i dont' like how it dirties the hosts mount 1170379020 M * daniel_hozac so, use mount -n. 1170379074 M * jjpacheco oh 1170379077 M * jjpacheco nice 1170379082 M * jjpacheco haha, thanks 1170379095 M * jjpacheco ok 1170379112 M * jjpacheco and what about writing the change to the vservers mtab 1170379137 M * daniel_hozac you'd need to use secure-mount for that. 1170379235 M * jjpacheco will i need to use -n with that also? 1170379252 M * daniel_hozac wouldn't make much sense when you're pointing it at the guest's mtab, would it? 1170379368 M * jjpacheco ah 1170379372 M * jjpacheco very true 1170379592 J * FireEgl ~FireEgl@adsl-63-199-202.bhm.bellsouth.net 1170379826 Q * Piet Quit: Piet 1170380518 M * jjpacheco daniel_hozac: secure-mount still adds the info to the hosts mtab and it doesn't show up in the vservers mtab 1170380528 M * jjpacheco basically, its the same thing as mount 1170380530 M * jjpacheco lol 1170380540 M * daniel_hozac jjpacheco: did you tell it to use the guest's mtab? 1170380568 M * jjpacheco how do i do that? 1170380599 M * daniel_hozac --mtab /vservers//etc/mtab? 1170380690 M * daniel_hozac it's not psychic, you know. 1170380827 M * jjpacheco haha, i guess its not 1170380970 M * jjpacheco ok 1170380975 M * jjpacheco so original question 1170380985 M * jjpacheco is what i originally wanted possible? 1170381008 M * jjpacheco add an entree to the guest fstab 1170381032 M * jjpacheco then issue a mount -a, and have the new addition take action 1170381223 M * daniel_hozac i don't see why not, but you'd need to use the right options for secure-mount. 1170381266 M * daniel_hozac i _really_ don't see what it gets you that just adding the mount yourself doesn't though. 1170381291 M * daniel_hozac (and i imagine there'd be duplicated mounts for the ones that already exist) 1170381462 M * jjpacheco well, i like the convenience of it all 1170381505 M * jjpacheco adding an entree to the guests fstab, then issuing a generic command is nice and simple 1170381538 M * jjpacheco why would i have duplicate mounts wiht mount -a? 1170381551 M * jjpacheco wouldn't it just re-mount everything in the vservers fstab? 1170381558 M * jjpacheco old and new 1170381585 M * daniel_hozac mount -a doesn't unmount anything. 1170381603 M * jjpacheco right 1170381615 M * jjpacheco i don't expect to unmount very often 1170381975 Q * duckx Remote host closed the connection 1170382031 M * jjpacheco ah, wicked 1170382035 M * jjpacheco got it working 1170382042 M * jjpacheco thx for the help daniel, u rock! 1170382685 M * jjpacheco one last question 1170382762 M * jjpacheco when i do the whole secure-mount thing (your method) the newly mounted fs has the full path listed in the guests mtab, how can i make it appear properly? 1170382799 M * daniel_hozac did you use the --chroot option? 1170382972 M * jjpacheco ah, niiiice 1170383401 M * Guy- http://linux-vserver.org/Running_runit-supervised_services_inside_a_vserver 1170383407 M * Guy- but now, sleep() 1170383708 J * adrien-modulis ~adrien@hvmoduli.enter-net.com 1170383902 M * jjpacheco i need the smallest guest possible. does anyone know how that might be accomplished? 1170383919 M * daniel_hozac smallest in what way? 1170383928 M * jjpacheco minimal 1170383932 M * jjpacheco very minimal 1170383947 M * jjpacheco im using debian, but its taking up 255M 1170383962 M * jjpacheco i'd like my guest to be around 50M or so 1170383981 M * daniel_hozac so, uninstall packages you don't need? 1170384004 M * jjpacheco i could do that 1170384015 M * jjpacheco but i was wondering if there was a package out there already 1170384048 M * daniel_hozac if you install (e.g.) an FC6 guest with the rpm build method, you'd get only glibc and what it requires (6 packages, IIRC). 1170384065 M * daniel_hozac that should be just a couple of mebibytes. 1170384087 M * jjpacheco would i be able to build one wiht the vserver build build command? 1170384089 M * daniel_hozac hmm, maybe not, seems one of the requirements is glibc-common which is rather large. 1170384102 M * daniel_hozac sure. 1170384116 M * jjpacheco i was thinking of damn small linux 1170384118 M * daniel_hozac you'd need to download the packages first though. 1170384123 M * jjpacheco its only 50M 1170384128 M * daniel_hozac compressed.... 1170384141 M * daniel_hozac try uncompressing that. 1170384146 M * jjpacheco but i couldn't find a bootstrap type installer (im doing this remotly) 1170384160 M * jjpacheco remotely 1170384230 M * jjpacheco i don't need like 95% of whats in these vservers 1170384241 M * jjpacheco the ability to ls 1170384242 M * jjpacheco cd 1170384245 M * jjpacheco rm 1170384247 M * jjpacheco vi 1170384259 M * jjpacheco chmod/chown/getacl/setacl 1170384260 M * jjpacheco etc 1170384263 M * jjpacheco very basic 1170384308 M * jjpacheco bah, i'll just stick with debian 1170384321 M * jjpacheco and try and cut down as much as possible 1170384400 Q * PowerKe Ping timeout: 480 seconds 1170385709 J * gerrit ~gerrit@c-67-160-146-170.hsd1.or.comcast.net 1170387012 J * PowerKe ~tom@d54C13E4B.access.telenet.be 1170387033 Q * PowerKe Server closed connection 1170387312 J * PowerKe ~tom@d54C13E4B.access.telenet.be 1170388372 J * Aiken_ ~james@ppp220-70.lns2.bne1.internode.on.net 1170388385 Q * Aiken Ping timeout: 480 seconds 1170390414 J * suka ~suka@bi01p1.co.us.ibm.com 1170390976 J * Aiken__ ~james@ppp220-70.lns2.bne1.internode.on.net 1170391014 P * suka Leaving 1170391295 Q * Aiken_ Ping timeout: 480 seconds 1170398588 Q * hallyn Ping timeout: 480 seconds 1170399553 Q * cdrx Ping timeout: 480 seconds 1170399847 J * gab ~gab@158.36.45.236 1170399950 N * Bertl_zZ Bertl 1170399955 M * Bertl morning folks! 1170400328 M * Bertl welcome gab! 1170400340 M * Bertl nebuchadnezzar: ping? 1170404630 M * Loki|muh Bertl: good morning! 1170404641 M * Bertl hey Loki|muh! LTNS! 1170404660 M * Loki|muh Bertl: have you heard about problems with ubuntu guests when starting them with --defaulttty? 1170404732 M * Loki|muh e.g. mysql and apache not starting 1170404748 Q * dlezcano Ping timeout: 480 seconds 1170404757 M * Bertl not yet, but I know that ubuntu as host is quite broken 1170404762 Q * kir Quit: Leaving 1170404775 M * Bertl Loki|muh: what does apache report as error? 1170404793 M * Loki|muh nothing ;) 1170404805 M * Bertl just dies away? 1170404822 M * Loki|muh just does not start, no entry in the log, nothing 1170404848 M * Loki|muh when started without --defaulttty, everythings fine 1170404882 M * Bertl to be honest, I do not even know the --defaulttty option 1170404887 M * Loki|muh daniel_hozac pointed me to try starting apache with < /dev/null &> /dev/null, which is causing the same problems 1170404909 M * Bertl well, that is definitely a start script bug then 1170404913 M * Loki|muh --defaulttty is used in /usr/lib/util-vserver/start-vservers 1170404920 M * Bertl probably in one of the startup functions ... 1170404958 M * Bertl a startup script failing when redirected is buggy in my book 1170405087 M * Bertl you might want to try apache without the wrapper and/or remove line by line from the wrapper 1170405107 M * Bertl especially the user feedback functions and includes might be causing this 1170405127 M * Bertl (e.g. by being too smart and trying to figure the screensize :) 1170405148 M * Loki|muh ok I will dive into that later and will report again ;) 1170405646 J * cdrx ~legoater@blueice4n2.uk.ibm.com 1170406774 M * Bertl okay, off for a second nap ... back later ... 1170406778 N * Bertl Bertl_zZ 1170407598 Q * click Ping timeout: 480 seconds 1170408751 M * eyck how well tested are we on itanium? 1170409175 J * dlezcano ~dlezcano@blueice2n1.uk.ibm.com 1170409465 Q * shedi Quit: Leaving 1170409465 J * derjoerg ~Miranda@199.42.240.136 1170409519 M * derjoerg hello everybody 1170409600 M * matti Morning. 1170409880 M * renihs morning 1170409925 M * derjoerg actually i'm running ubuntu dapper with kernel 2.6.17.6 with vs 2.1.1 rc29 from derjohn 1170410004 M * derjoerg yesterday I tried to compile vanilla kernel 2.6.19.2 with vs 2.2.0 rc9 1170410021 M * derjoerg everything went fine until I ran "lilo" 1170410043 M * derjoerg I got the following warning 1170410043 M * derjoerg Warning: '/proc/partitions' does not match '/dev' directory structure. 1170410043 M * derjoerg Name change: '/dev/dm-0' -> '/dev/evms/.nodes/sda1' 1170410060 M * derjoerg and after a reboot the system did not came up 1170410096 M * derjoerg I had to start the rescue-system of my server-provider and had to reenable the old kernel, than everything worked fine 1170410131 M * derjoerg for information I use a softraid 1 1170410149 M * derjoerg anybody out there with some tips/tricks? 1170410233 M * waldi derjoerg: you use evms, and don't know how to boot it is the dist provided scripts don#t do their work 1170410234 J * dna ~naucki@90-249-dsl.kielnet.net 1170410391 M * derjoerg waldi: sorry, I'm a real newbie for all of this and don't really understand what you mean 1170410425 M * waldi exactly, you don't know what you do 1170410491 M * waldi read the evms documentation, how to boot with it enabled 1170410508 M * derjoerg derjoerg: what do you want to say. I want to learn. Do you mean "You have no glue what to do, so I will not help you!" ? 1170410525 M * derjoerg waldi: ok 1170410590 M * derjoerg I'm only a little bit confused, why the boot works with kernel 2.6.17.6 and not with 2.6.19.2 1170410625 M * waldi show the console output 1170410692 J * DavidS ~david@vpn.uni-ak.ac.at 1170411081 M * derjohn derjoerg, evms ... interesting ... waldi i observed ubuntu assembeling _all_ softraids at boot time. the first n lines of the console log at boot from the kernel is "assemble raid blah". I never spotted this behavior from a vanilla or debian kernel. Had that to to with evms? Is it a kernel feature that can be enabled in .config? or maybe a ubuntu addon? 1170411122 M * waldi evms, AFAIK 1170411136 M * waldi i dropped that piece of software long time ago 1170411189 M * derjohn in favor of lvm ? 1170411255 M * derjoerg waldi: the server is hosted by a provider, who will be paid, that I have console login 1170411294 M * derjohn derjoerg, maybe we should check ubuntus .config and see if they compiled evms hard into the kernel ... 1170411336 M * derjoerg derjohn: I used the config from your kernel 2.6.17.6 with vs 2.1.1 rc29 1170411441 M * derjohn well, it in /boot/config-fooXX 1170411446 M * derjohn *it's 1170411723 M * derjoerg I used /boot/config-2.6.17.6-ubuntu1 1170411874 M * derjoerg can I check somewhere in "make menuconfig" if evms is enabled? 1170412007 M * dm8tbr derjoerg: when building a kernel i generally use the previous config for the new sources and run "make oldconfig" to get asked the new options 1170412088 M * derjoerg dm8tbr: sure, I've done this with the config from kernel 2.6.17.6, but there where a lot of additional questions, for which I simple pressed "enter" 1170412185 M * dm8tbr also this won't tell you if the older kernel was patched for something. I had this when i forgot to apply the vesafb-tng patch... no console output at all... 1170412431 M * derjoerg dm8tbf: Is there a way, how I can figure this out? 1170412577 M * dm8tbr Best way should be to check out the patchset for the older kernel 1170412584 M * dm8tbr ubuntu provides those 1170413141 M * derjoerg do you have a clue, where I can found them? I don't find actually something 1170413374 M * dm8tbr derjoerg: http://packages.ubuntu.com/edgy/source/linux-source-2.6.17 1170413514 Q * derjoerg Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org 1170413515 M * dm8tbr there's probably a even more elegant way than getting 2M of gzipped diff. but i don't know where i'd find a patch list 1170413668 M * Guy- what could be wrong if vserver start says vshelper.init: can not determine xid of vserver 'maildino'; returned value was ''? 1170413801 M * Guy- OK, probably a problem with init in the vserver 1170414259 Q * Aiken__ Quit: Leaving 1170414416 M * renihs hmm on the host system can i leave /etc/make.profile -> ../usr/portage/profiles/default-linux/x86/2006.1 or should i use a server profile? 1170414449 M * daniel_hozac AFAIK the host's profile doesn't matter. 1170414922 J * Aiken ~james@ppp220-70.lns2.bne1.internode.on.net 1170415539 M * mugwump does anyone have a demo program using the new clone flags? 1170415734 M * daniel_hozac util-vserver? (though admittedly, that uses them indirectly via the kernel) 1170415789 M * daniel_hozac there really isn't anything more to it than unshare(CLONE_NEWUTS); execvp(argv[1], argv+1); though. 1170415792 M * daniel_hozac (e.g.) 1170416584 J * ciphergoth ~paul@host226.lshift.net 1170416666 Q * Aiken Quit: Leaving 1170417019 J * duckx ~Duck@tox.dyndns.org 1170417178 J * ziga ziga@al.co.ve 1170417181 M * ziga Hello 1170417252 M * ziga I'm trying to use mount --bind to share a directory through multiple vservers, it has been working in the past, but it seems that now, files created within a vserver aren't shared with others, not even with the main context (though, files created within the main context are shared across all vservers) 1170417287 M * ziga Anyone knows how to fix this behavior ? ;-) 1170417300 M * daniel_hozac is the filesystem you're bind mounting mounted with tagxid? 1170417368 J * kir ~kir@swsoft-mipt-nat.sw.ru 1170417387 M * ziga I did not specify any option :s 1170417401 M * ziga /dev/sda9 218G 1.8G 205G 1% /vservers 1170417419 M * ziga err 1170417420 M * ziga /dev/sda9 /vservers ext3 defaults 0 2 1170417586 M * ciphergoth If the vserver is stopped, can I freely modify the filesystem in /var/lib/vservers? 1170417919 M * daniel_hozac ciphergoth: sure. 1170417931 M * ciphergoth cool 1170417972 M * daniel_hozac ziga: so you're sure the bind mount is in place and doing what you expect it to? can you see files there which are supposed to be there? 1170417978 M * daniel_hozac (in all of the guests) 1170417981 J * rpc ~no@ahm234.neoplus.adsl.tpnet.pl 1170417987 M * rpc hello everyone 1170417989 M * daniel_hozac ziga: also, what kernel is this? 1170417992 M * daniel_hozac hello rpc 1170418043 M * rpc nice to meet you daniel 1170418123 M * rpc this is my very first time with vserver and before i get down to work with installation i would like to make sure if 2.6.8-12-amd64-k8-smp is fine for that? 1170418159 M * rpc not really sure where to get a vserver patch for this kernel 1170418167 M * daniel_hozac no. 1170418184 M * daniel_hozac i'd suggest you get a vanilla 2.6.19.2 kernel, and use the 2.2.0-rc9 patch. 1170418200 M * rpc oh hm 1170418231 M * rpc thanks very much, this hint saved me alot of time already 1170418310 M * ex daniel_hozac, 2.2.0-rc9 is stable enough for production servers? 1170418313 M * rpc but all in all i shouldn't expect problems running vserver on dual core amd64, right? 1170418323 M * rpc debian64 as well 1170418369 M * daniel_hozac ex: we sure hope so, as that's the next stable branch, with 2.0 dying. 1170418396 M * daniel_hozac rpc: i'm running Fedora on my dual core AMD64, works just fine. 1170418410 M * rpc sounds great :) 1170418411 J * lilalinux ~plasma@dslb-084-058-221-110.pools.arcor-ip.net 1170418469 M * rpc ok, so the first step - vanilla 2.6.19 with smp and next step 2.2.0-rc9 patch 1170419287 M * mEDI_S hmm includet rc9 a fix for the 64bit problem in the stats from /proc/virtual/$ID/cacct ? 1170419721 M * daniel_hozac no. 1170419741 M * daniel_hozac but i should have a patch for that later today... 1170419816 M * mEDI_S kk thx 1170420734 Q * pusling Read error: Connection reset by peer 1170421097 J * shedi ~siggi@dsl-149-109-85.hive.is 1170421805 M * renihs hmm gentoo howto describes a vserver-new script which i seem to be missing, can i use the stage3 file like normal and create all required stuff in /etc/vservers/ without it? 1170421816 M * renihs the vserver stage3 template i mean 1170421847 M * daniel_hozac use vserver ... build -m template --context... -- -t -d gentoo 1170421942 M * renihs mkay 1170421975 M * ex linux-vserver.org server seems to have some problems with mysql 1170421976 M * renihs -d ? hmm 1170421984 M * renihs is that init style? 1170421999 M * daniel_hozac no, that's distribution. 1170422006 M * renihs mkay 1170422009 M * renihs thanks 1170422032 J * pusling pusling@195.215.29.124 1170422861 Q * q\ Read error: Connection reset by peer 1170422877 J * q\ ~java@avr49-1-82-245-33-193.fbx.proxad.net 1170422908 M * ziga daniel_hozac: yup ! I'm sure 1170422909 M * ziga daniel_hozac: Linux mk2-sde-lvs1 2.6.17.13-vs2.0.2.1supermicro #1 SMP Fri Jan 26 14:11:51 CET 2007 i686 GNU/Linux 1170423047 M * nox *sigh* my noobquestion of the day: how to shutdown an interface when ip addr it still shows to be up after stopping the guest? 1170423084 M * daniel_hozac ip addr del dev ? 1170423090 M * nox scope global eth0 <-- do not want to shutdown eth0 1170423110 M * nox ah ok i try 1170423185 M * nox thx did it 1170423335 Q * kaner_ Remote host closed the connection 1170423742 J * kaner kaner@strace.org 1170424092 J * dreamind ~dreamind@C2107.campino.wh.tu-darmstadt.de 1170424096 Q * dreamind 1170424101 Q * mnemoc Ping timeout: 480 seconds 1170424271 Q * m`m`h Read error: Connection reset by peer 1170424430 J * mnemoc ~amery@kilo105.server4you.de 1170424958 Q * shedi Ping timeout: 480 seconds 1170425070 J * _dmax ~semaj@bl9-225-238.dsl.telepac.pt 1170425141 Q * dmax Ping timeout: 480 seconds 1170425141 N * _dmax dmax 1170425223 J * m`m`h ~simba@deb30.mgts.by 1170425650 J * shedi ~siggi@dsl-149-109-85.hive.is 1170426080 Q * mnemoc Ping timeout: 480 seconds 1170426092 J * mnemoc ~amery@kilo105.server4you.de 1170426259 J * hallyn ~xa@cpe-72-179-43-119.austin.res.rr.com 1170427890 J * virtuoso_ ~s0t0na@shisha.spb.ru 1170427899 Q * virtuoso cation.oftc.net kinetic.oftc.net 1170428505 Q * shedi Quit: Leaving 1170429319 J * ybanafa ybanafa@89.189.67.235 1170429658 J * NextSuX ~nexus@office.netbox.cz 1170429724 Q * ybanafa Remote host closed the connection 1170429812 M * NextSuX Hi, I'm relatively new to vserver enviroment. I have problem with setting SHMMAX in virtual server context. I've been used vserver inherited this value from host, but now under 2.6.19 + vs2.1.1.7 it somehow does not... I've been googling for hours but with no result :( (except one thread with no solution). Can you give me please any hint? 1170430079 Q * nox Remote host closed the connection 1170430119 M * daniel_hozac that's set in proc, right? 1170430136 M * NextSuX daniel_hozac: well.. in normal linux yes 1170430155 M * NextSuX but now: 1170430156 M * NextSuX echo 3000 > /proc/sys/kernel/shmmax 1170430156 M * NextSuX bash: /proc/sys/kernel/shmmax: Operation not permitted 1170430171 M * NextSuX same result with sysctl 1170430202 M * NextSuX I've tried rlimits which was suggested in mailing list archive, but it did not work 1170430231 M * NextSuX cat /proc/virtual/244/limit | grep SHM 1170430234 M * daniel_hozac with util-vserver 0.30.213-pre3+, you should be able to do echo kernel.shmmax > /etc/vservers//sysctl/0/setting; echo 3000 > /etc/vservers//sysctl/0/value 1170430240 M * NextSuX SHM: 8192 0/ 8192 3000/3000 0 1170430258 M * daniel_hozac (after creating the /etc/vservers//sysctl/0 directory, of course) 1170430269 M * NextSuX hmh... this version is not in main gentoo portage - but I'll try 1170430315 M * daniel_hozac btw, you probably should use 2.2.0-rc9. we haven't updated 2.1 for a while... 1170430445 M * daniel_hozac ziga: so how do you do the bind mounts? with the fstab in the guest configuration? 1170430676 M * NextSuX daniel_hozac: where can I find 0.30.213-pre3+ version? I can't find it anywhere :( 1170430700 M * daniel_hozac http://people.linux-vserver.org/~dhozac/p/uv/experimental/util-vserver-0.30.213-rc1.tar.bz2 1170430716 M * NextSuX thanks 1170430957 J * nox ~nox@static.88-198-17-175.clients.your-server.de 1170431587 P * adrien-modulis 1170431837 M * NextSuX daniel_hozac: AWESOME MAN!! It works - thanks a LOT! 1170432109 M * ziga daniel_hozac: in the host configuration 1170432589 M * ciphergoth debian host, fedora guest: "su mysql" works, but MySQL complains "Fatal error: Can't change to run as user 'mysql'.1 1170432592 M * ciphergoth Any ideas? 1170432628 M * daniel_hozac nothing more descriptive? 1170432690 M * daniel_hozac ziga: meaning the host's /etc/fstab? 1170432708 M * ciphergoth The only useful information in /var/lib/mysql/glue2.arizonabay.lshift.net.err reads 1170432713 M * ciphergoth Fatal error: Can't change to run as user 'mysql'. Please check that the user exists! 1170432734 M * ciphergoth oh sorry, thought you meant me.,.. 1170432833 M * ciphergoth strange, I don't have an /etc/my.cnf or similar 1170432908 M * ziga daniel_hozac: yep 1170432967 M * ciphergoth oh great. rpm says MySQL-server-4.0.27-0.i386.rpm is not installed when I -e it but that it's already installed when I -i it. 1170433005 M * ciphergoth ignore that last 1170433410 Q * NextSuX Quit: Leaving 1170433706 Q * DavidS Quit: Leaving. 1170434004 J * stefani ~stefani@flute.radonc.washington.edu 1170434024 M * daniel_hozac ciphergoth: might want to strace that to potentially get some more info. 1170434396 M * dm8tbr daniel_hozac: do you happen to know if the v6 patch against 2.6.19.1 will apply against 2.6.19.2 or if bonbons is working on that? 1170434995 M * ciphergoth I think my Fedora is missing some essential packages eg nscd - how do I get a list of what it should have? 1170435085 M * ciphergoth something is trying to connect to /var/run/.nscd_socket rather than /var/run/nscd/socket 1170435140 M * daniel_hozac dm8tbr: it should apply. 1170435161 M * daniel_hozac dm8tbr: if it doesn't, bonbons or i can probably fix that up rather quick. 1170435204 M * dm8tbr daniel_hozac: thank you very much for the info! 1170435229 M * daniel_hozac ciphergoth: nscd shouldn't be essential. 1170435257 M * ciphergoth mysql seems to be making no other attempt to find out the mysql uid 1170435267 M * ciphergoth it doesn't open /etc/passwd or anything 1170435359 M * daniel_hozac where did you get the MySQL packages from? mysql.com? 1170435380 M * ciphergoth good question 1170435395 M * daniel_hozac because the Fedora packages are named mysql-server, without the caps. 1170435396 M * ciphergoth I'm trying to install the exact same MySQL I found on another server 1170435401 M * ciphergoth 4.0.27 1170435444 M * ciphergoth man 1170435456 M * ciphergoth even mysql won't let me download that even in their archives 1170435469 M * ciphergoth just possibly I should use a slightly more recent version 1170435484 M * daniel_hozac just yum install mysql-server.. 1170435583 M * ciphergoth that'll be far too new - 5.something 1170435596 M * ciphergoth 4.1.something is probably the best compromise 1170435663 M * daniel_hozac ziga: so could you explain exactly how to reproduce your problem? e.g. what would i need to do to get your results? 1170435670 M * dm8tbr daniel_hozac: friend of mine tried it - http://nopaste.biz/?12034 :/ 1170436020 M * daniel_hozac http://people.linux-vserver.org/~dhozac/p/k/patch-2.6.19.2-vs2.2.0-rc9.ipv6.diff 1170436029 M * daniel_hozac completely untested, but it should fix those conflicts. 1170436080 Q * dlezcano Read error: Connection reset by peer 1170436088 Q * FireEgl Quit: ... 1170436125 M * dm8tbr tnx 1170436160 J * Piet hiddenserv@tor.noreply.org 1170436176 Q * PowerKe Ping timeout: 480 seconds 1170437076 Q * id23 Ping timeout: 480 seconds 1170437283 Q * cdrx Read error: Connection reset by peer 1170437580 J * id23 ~id@p50810B48.dip0.t-ipconnect.de 1170437759 Q * ciphergoth Quit: Client exiting 1170438485 N * Bertl_zZ Bertl 1170438490 M * Bertl morning folks! 1170438540 J * the_hydra ~a_mulyadi@125.164.96.178 1170438644 N * jbailey-afk jbailey 1170438654 M * Bertl welcome the_hydra! jbailey! 1170438680 M * rpc daniel_hozac: am i safe with this kernel? linux-image-2.6.18-3-xen-vserver-686 1170438681 M * daniel_hozac morning Bertl! 1170438688 M * daniel_hozac rpc: no, that's known to be broken. 1170438693 M * daniel_hozac rpc: you should use at least 2.6.18-4. 1170438695 M * rpc oh damn it 1170438704 M * rpc i spent all day fixing the box 1170438705 M * jbailey Bertl: Heya. I'm still here from a couple days ago. =) 1170438711 M * rpc and now i'm broken again :/ 1170438736 M * daniel_hozac Bertl: is the princeton machine up for you? i'm getting connection refused. 1170438764 M * Bertl sec, checking 1170438851 M * Bertl daniel_hozac: works here ... 1170438910 M * the_hydra Bertl: hi , sorry, having a very very terrible lag here 1170439062 J * comfrey_ ~comfrey@70.91.185.84 1170439098 J * comfrey_1 ~comfrey@70.91.185.84 1170439171 Q * comfrey Ping timeout: 480 seconds 1170439186 M * ziga daniel_hozac: Ok, it's easy: boostrap two vservers, on the host: mkdir /tmp/testdir; mkdir /vservers/a/tmp/testdir; mkdir /vservers/b/tmp/testdir; mount -t union --bind /tmp/testdir /vservers/a/tmp/testdir; mount -t union --bind /tmp/testdir /vservers/b/tmp/testdir; vserver a enter, touch /tmp/testdir/testfile, vserver b enter, ls /tmp/testdir 1170439203 Q * comfrey__ Ping timeout: 480 seconds 1170439208 M * daniel_hozac -t union? 1170439228 M * Bertl ziga: unionfs is not in mainline 1170439262 M * Bertl ziga: did you patch it yourself? 1170439310 M * daniel_hozac having never looked at unionfs, does it intercept --bind mounts and do its own thing? 1170439317 M * ziga Hrm wait, I'm wondering right now if I used unionfs or not 1170439323 M * ziga I'm checking ;-) 1170439351 M * Bertl daniel_hozac: unionfs doesn't cope with namespaces properly (yet) 1170439368 M * Bertl daniel_hozac: I guess it will (have to) when it actually gets included 1170439378 M * ziga /vservers/mk2-lvs-nas1/storage on /vservers/mk2-lvs-encoding1/storage type union (rw,bind) 1170439380 M * daniel_hozac yeah, right... 1170439381 M * ziga /vservers/mk2-lvs-nas1/storage on /vservers/mk2-lvs-jboss1/storage type union (rw,bind) 1170439384 M * ziga ! 1170439395 Q * the_hydra Ping timeout: 480 seconds 1170439400 M * ziga Alright .. 1170439405 M * daniel_hozac ziga: is that from mount? or /proc/mounts? 1170439411 M * ziga mount 1170439414 M * daniel_hozac mount lies. 1170439441 M * Bertl well, it seems to be a bind mount then :) 1170439457 M * Bertl with a funny -t union added :) 1170439486 M * Bertl ziga: I assume you didn't patch unionfs into the kernel, right? 1170439508 M * ziga Bertl: I didn't 1170439530 M * ziga Though I can try if you believe it's a bug ;-) 1170439533 M * Bertl okay, so that clarifies, please remove the -t union from your example above, as it is just confusing :) 1170439548 M * Bertl no, I don't think it will actually change anything 1170439568 M * Bertl you problem is a different one 1170439570 M * ziga Lemme backup the stuff inside, and I try again :-) Gimme 5 minutes 1170439585 M * Bertl you are trying to mount a dir into two guests for data sharing, right? 1170439604 M * ziga yup 1170439626 J * shedi ~siggi@ftth-237-144.hive.is 1170439671 M * ziga /vservers/mk2-lvs-nas1/storage on /vservers/mk2-lvs-jboss1/storage type none (rw,bind) 1170439672 M * Bertl ziga: okay, your experiment with guest /tmp/dir probably fails, because it is replaced with the /tmp mounted into each guest 1170439674 M * ziga /vservers/mk2-lvs-nas1/storage on /vservers/mk2-lvs-encoding1/storage type none (rw,bind) 1170439694 M * Bertl (assumed you start the guest after doing the --bind mount) 1170439728 M * Bertl this example, with the storage, should work, if you start the guests after doing that, and keep them from removing the mount 1170439741 M * ziga Hehe. Well, I did not test with /tmp in fact, I'm restarting the vservers right now .. 1170439758 M * Bertl daniel_hozac: how would ziga prevent the cleanup for this path? 1170439771 M * daniel_hozac is it necessary? 1170439782 M * daniel_hozac we only need the mount inside the guest to live on, right? 1170439785 M * daniel_hozac those are already excluded. 1170439786 M * Bertl no idea, you do the userspace cleanup :) 1170439824 M * Bertl i.e. if you say, it isn't necessary to exclude it anywhere, so be it :) 1170439832 M * ziga Bertl, daniel_hozac: It works now ;-) Thanks ! 1170439840 M * Bertl you're welcome! 1170439840 M * daniel_hozac what did you change? 1170439851 M * ziga I had to remove that -t union (don't ask me why I did that) 1170439862 M * ziga And I also probably did not restart the vservers after remounting the partitions 1170439873 M * Bertl btw, you can make that a little more natural by using a common dir, and mounting that into the guests like this: 1170439886 M * Bertl dir /storage on the host 1170439897 M * Bertl entry (in fstab for each guest config) 1170439910 M * ziga Well, 1170439920 M * ziga /dev/sda9 218G 1.8G 205G 1% /vservers 1170439920 M * Bertl /storage /storage none bind 1170439952 M * Bertl which will make a bind mount on guest startup 1170439973 M * Bertl from /storage to /vservers//storage when the guest is started 1170439993 M * Bertl and which will also reove that mount when the guest is stopped again 1170439999 M * ziga Ahh, Right 1170440258 J * cdrx ~legoater@cap31-3-82-227-199-249.fbx.proxad.net 1170440313 M * harry hahaaaaaaaa 1170440326 A * harry interdiffs 2.2.0-rc[89] 1170440331 M * harry well... not now, but soon! :) 1170440622 M * Bertl with a little luck, we have a final release with rc9, daniel_hozac what's your opinion? 1170440642 M * daniel_hozac longatomic and http://people.linux-vserver.org/~dhozac/p/k/delta-ptrsize-fix01.diff ? 1170440660 M * daniel_hozac (the longatomic i'm about to test is http://people.linux-vserver.org/~dhozac/p/k/delta-longatomic-fix01.diff) 1170440676 Q * Piet Ping timeout: 480 seconds 1170440683 M * daniel_hozac Bertl: does the grub on there support booting another entry just once? 1170440709 M * Bertl just make an entry, and tell me which one to boot 1170440723 M * Bertl ah, and let me close my remote apps first :) 1170440728 M * daniel_hozac the linux-2619-vs2209l entry 1170440827 M * Bertl and yes, it supports one time booting, but in a strage way, mandriva folks decided to implement -R in grub?? 1170440853 M * daniel_hozac heh, i guess every distro has their own way... 1170440854 M * Bertl i.e. I have to wipe that out and install a normal grub there :) 1170440879 J * PowerKe ~tom@d54C13E4B.access.telenet.be 1170440935 M * dm8tbr anyone got an idea how to fix disk space accounting not working? it always says rootfs 5120000 0 4860000 0% / 1170440949 M * daniel_hozac are you using disk limits? 1170441004 M * dm8tbr it is not my machine, i'm just using one of the vservers. the admin made me ask here ;) 1170441013 M * dm8tbr I'll ask him 1170441025 M * daniel_hozac would probably be easier if he just came on here himself ;) 1170441046 M * dm8tbr yea, same thought same moment 1170441398 J * baldy 1001@83.223.66.130 1170441399 M * baldy huhu 1170441403 M * baldy ahh works... 1170441461 M * baldy daniel_hozac: i dont use it... 1170441473 M * dm8tbr ah, there he is :) 1170441478 M * baldy i make all the shit it openvcp ;) 1170441485 M * baldy with.. 1170441504 M * daniel_hozac well, i've never used nor seriously looked at openvcp's code... 1170441548 M * daniel_hozac that sure looks like disk limits on a filesystem without tagxid though. 1170441564 M * baldy funny is... that the vps are allwasa dev/hdv1 or is it normal? 1170441572 M * baldy i dont have any limits... 1170441599 M * daniel_hozac yes, the default mtab for guests uses /dev/hdv1 for the root mount. 1170441676 M * baldy how can i set a limit for the vServers? 1170441719 M * Bertl you need to have xid tagging enabled for that to work 1170441860 M * baldy in the kernel? 1170441890 M * Bertl in the kernel, and in userspace, when you do the mount 1170441921 M * baldy that i can do hehe 1170441951 M * baldy u mount each vps as a own partition? 1170441973 M * Bertl well, you can do that too, in this case you do not need the tagging at all 1170441986 M * Bertl but the idea is to have limits on a shared partition per guest 1170441997 M * baldy yep thats what i want ;) 1170442132 M * Bertl then you put a bunch of guests on a shared partition, with tagging enabled 1170442149 M * Bertl and specify the desired limits (inodes and space) in the vserver config 1170442169 M * Bertl on guest startup, the actually used space will be calculated, and the kernel will keep track of it 1170442267 M * daniel_hozac Bertl: /proc/virtual/42/cacct looks ok, right? 1170442296 Q * shuri Remote host closed the connection 1170442347 M * Bertl yep 1170442355 M * Bertl (had to install bc :) 1170442391 M * daniel_hozac hehe 1170442396 M * Bertl good work .. will integrate the two patches ... 1170442417 M * Bertl so I guess we do an rc10 and run it through plm them 1170442429 M * daniel_hozac sounds good. 1170442543 J * Piet hiddenserv@tor.noreply.org 1170442582 M * daniel_hozac (http://people.linux-vserver.org/~dhozac/t/tests/test-cacct.c is what i used to test, btw) 1170442633 M * Bertl okay, great! 1170443031 J * click click@ti511110a080-0186.bb.online.no 1170443062 M * Bertl welcome click! LTNS! 1170443345 Q * baldy Read error: Connection reset by peer 1170443363 Q * gerrit Ping timeout: 480 seconds 1170443813 Q * Hollow Remote host closed the connection 1170443833 J * Hollow ~hollow@styx.xnull.de 1170443849 M * daniel_hozac Bertl: i get lots of lib/vserver-internal.h:260: warning: initialization makes integer from pointer without a cast 1170443856 M * daniel_hozac with shiny14. 1170443868 M * daniel_hozac (line 260 of lib/vserver-internal.h is the _syscall3 line) 1170444137 M * Bertl hmm ... 1170444186 M * Bertl is that a call passing a pointer? 1170444195 M * daniel_hozac it's the vserver call, so yeah. 1170444238 M * Bertl okay, probably a casting is wrong then, could you run it through cpp and try that one? 1170444243 M * Bertl so that we see the actual line? 1170444402 M * Bertl might be worth running the .E through indent -kr first :) 1170444434 J * suka ~suka@bi01p1.co.us.ibm.com 1170444448 M * suka Bertl: ping 1170444514 M * Bertl suka: pong! 1170444531 M * daniel_hozac there's no (long) before the arguments. 1170444531 M * suka Bertl: my mails to containers list are being intercepted by Smartlist and require me to verify my address 1170444551 M * suka I verified it yday, but for another mail today, I got the same resp 1170444551 M * Bertl when assigning a register or what? 1170444552 Q * lilalinux Remote host closed the connection 1170444559 M * daniel_hozac i haven't verified it, but i assume register long __sc_a3 __asm__ ("rdx") = (data); is the problem. 1170444579 M * Bertl suka: what am I supposed to do? 1170444590 M * daniel_hozac suka: i assume you mean the vserver list? 1170444599 M * suka The mail is intercepted from vserver list 1170444606 M * suka yes 1170444614 M * Bertl containers or vserver? 1170444628 M * suka I send the mail to containers, but get the intercept for vserver 1170444643 M * daniel_hozac suka: is your From: the same? 1170444649 M * Bertl that just because the container folks probably didn't get the subscription right 1170444666 M * Bertl (or they have no valid reverse, etc ...) 1170444676 M * daniel_hozac ah, the vserver list is subscribed to the containers list? 1170444686 M * Bertl it should be by now, but I doubt it 1170444694 M * suka I guess vserver@list.linux-vserver.org is subscribed to containers 1170444705 M * suka yes 1170444727 M * Bertl so most likely whatever is distributing the email, has a bad setup 1170444754 M * suka This seems to have started yday 1170444760 M * Bertl don't worry too much about it, if you want it to reach the vserver ML, subscribe there ... 1170444785 M * daniel_hozac (and probably CC the list too) 1170444789 M * Bertl alternatively, you can contact the respective ML maintainers 1170444812 M * Bertl in case of the vserver list, this would be Martin 1170444845 M * Bertl for the containers list, probably somebody by IBM? cdrx maybe? 1170444916 N * ebiederm_away ebiederm 1170444961 M * Bertl daniel_hozac: ah, right, I see it now 1170444984 M * daniel_hozac __sc_areg needs (__sysc_type), right? 1170445003 M * ebiederm I think I saw Dave Hansen say something about the containers list. 1170445013 M * daniel_hozac that seems to have silenced gcc, at least. 1170445033 M * suka So you think its a containers list issue ? 1170445038 M * ebiederm I do know that the problem suka say I have seen another complaint from one of the developers on netdev. 1170445066 M * Bertl daniel_hozac: yes, just checking for other palces too 1170445068 M * ebiederm suka: I think it is an issue with how the vservers-list is subscribed to containers. 1170445069 M * suka ebiederm: do you get a similar response from vserver when you post to containers ? 1170445075 M * Bertl daniel_hozac: will update to a 14b shortly 1170445087 M * daniel_hozac Bertl: okay. 1170445089 M * ebiederm suka: Stephen Hemminger did. 1170445175 M * ebiederm suka: So the solution needs to be either the vservers-list accepts everything from the containers-list, or we unsubscribe the vservers-list, from containers. 1170445219 M * Bertl please contact Martin for the former, but I do not really care 1170445233 M * suka ebiederm: ok. 1170445241 M * suka Bertl: Can you give me Martin's email 1170445254 M * Bertl yes, sec 1170445289 M * suka It would be best if we have one list (containers) for all. Or we will have to remember to cc vserver 1170445299 M * Bertl Martin List-Petersen 1170445321 M * ebiederm Bertl: The practical problem is that if the vservers list filters the content people won't want to review patches cc'd to containers. 1170445333 M * Bertl well, we didn't get a single containers email to the vserver list yet, so I guess it's fine if you keep it on containers 1170445361 M * daniel_hozac i.e. unsubscribe the list? 1170445363 M * Bertl ebiederm: there should be no content filtering except for obvious spam 1170445364 M * hallyn well thats a problem 1170445373 M * daniel_hozac Bertl: there's the greylisting. 1170445377 M * ebiederm Bertl: Really I hadn't looked so I guess the subscription is totally messed up. 1170445387 M * Bertl daniel_hozac: well, graylisting should be fine too 1170445406 M * Bertl daniel_hozac: otherwise I would not get a single mail either 1170445442 M * Bertl hallyn, ebiederm: I reported it several times, but gave up on some point ... 1170445516 M * ebiederm Then I suspect the path of least resistance is to just unsubscribe the vserver list then. 1170445520 M * Bertl daniel_hozac: okay, updated the shiny14 in place, should now do the proper thing 1170445559 M * Bertl ebiederm: no objection here ... 1170445607 M * suka Bertl: are you separately subscribed to containers ? 1170445616 M * Bertl yes, I am 1170445668 M * Bertl but in general I prefer to get CC-ed if you want me to read it in a timely manner ... 1170445737 J * dhansen ~dave@pool-72-90-117-15.ptldor.fios.verizon.net 1170445748 M * Bertl welcome dhansen! 1170445751 M * dhansen Bertl: Hi 1170445767 M * dhansen suka: said you thought there might be something wrong with the containers@ list 1170445803 M * Bertl no, I think there is something wrong with the subscription to the vserver list, or with the host distributing the emails of the containers list 1170445812 M * daniel_hozac Bertl: starting, entering and stopping a guest works fine with shiny14. is there any specific tests you'd want me to do? 1170445819 M * daniel_hozac s/is/are/ 1170445822 M * ebiederm dhansen: Having the vservers list subscribed is causing people to get weird bounce messages. 1170445837 M * Bertl I guess we should find a syscall with 6 args, socket stuff comes to my mind? 1170445856 Q * Piet Ping timeout: 480 seconds 1170445877 M * dhansen ebiederm: should we just pull it off then? Are enough people on containers@ now? 1170445909 M * Bertl dhansen: until now, I didn't see a single email to reach vserver yet 1170445909 M * ebiederm dhansen: I think so, and the vservers list has not gotten a single containers@ post. 1170445941 M * dhansen OK, I'll pull it off 1170445989 M * ebiederm Thanks. Then I won't get any more complaints when I post to netdev and cc containers :) 1170446254 M * Bertl daniel_hozac: hum, looking at vx_sock_count/total, I wonder why your patch actually works :) 1170446357 M * daniel_hozac oh? 1170446362 M * daniel_hozac did i miss something? 1170446445 M * Bertl include/linux/vserver/cacct_int.h says atomic_read() :) 1170446450 M * daniel_hozac ah, yeah.... 1170446452 M * daniel_hozac hum. 1170446476 M * Bertl changed it here, so it will be fixed in rc10 1170446481 M * daniel_hozac thanks. 1170446495 M * Bertl I also changed the (yet unused) dump to use lu 1170446540 M * daniel_hozac yeah, the %d seemed wrong to me, but i didn't do anything about it. 1170446580 Q * duckx Remote host closed the connection 1170446991 J * Piet hiddenserv@tor.noreply.org 1170447066 M * Bertl welcome Piet! 1170447093 M * Bertl ebiederm, hallyn: regarding the hook permission (older email) 1170447109 J * DotHack ~chatzilla@cp676338-a.landg1.lb.home.nl 1170447113 M * Bertl I think special care has to be taken there 1170447114 M * DotHack Hi all 1170447120 M * Bertl welcome DotHack! 1170447160 M * Bertl ebiederm, hallyn: the CAP_DAC_OVERRIDE should allow for guest root, but not for context manipulations, or did I musunderstand that? 1170447179 M * DotHack i have a tried to reboot a vserver with reboot -f inside it. now i got stuck with a vserver running 1 process wich is state D. i cant kill the process and i cant restart the vserver, can any one help me? 1170447209 M * Bertl D state means it either does something with your disk, or you got a kernel oops, check with dmesg 1170447230 M * Bertl (D is device I/O) 1170447239 M * DotHack noop got no dmesg 1170447265 M * daniel_hozac what process is that? 1170447272 M * Bertl what kernel is that? (version and maybe distribution) 1170447295 M * DotHack wait 1170447298 M * ebiederm Bertl: I agree that special care has to be taken. The capabilities don't don't fit in well anymore. 1170447327 M * Bertl ebiederm: I'd suggest to have a CAP_SPACE or so 1170447341 M * DotHack vmlinuz-2.6.18-3-vserver-686 1170447349 M * Bertl debian? 1170447356 M * DotHack yep 1170447365 M * Bertl that is known to be broken for some time now 1170447372 M * Bertl and you have a message in dmesg :) 1170447389 M * Bertl please upgrade to at least 2.6.18-4 and leave a message on the bugtracker 1170447417 M * ebiederm Bertl: Capabilities in linux have never fit well. for the capabilities that connect to namespaces the only clean bit feels like having a separate instance of the capability for each namespace. 1170447422 M * DotHack yeah i replced the kernel already with this: vmlinuz-2.6.18.5-vs2.1.1.3-p3-squash-drbd-256ip-ipv6 1170447454 M * Bertl ebiederm: we use a capability mask (in case you don't know) which works quite fine ... 1170447456 J * juggo ~lemur@h-68-166-181-4.sttnwaho.covad.net 1170447462 M * Bertl welcome juggo! 1170447471 M * juggo hey how do I kill processes in a vserver from the host? 1170447475 M * DotHack i got errors when shutting down vserver when 1170447479 M * Bertl juggo: vkill 1170447492 M * DotHack with the debian stock kernel 1170447494 M * juggo yeah was trying that, maybe I have the syntax wrong? 1170447504 M * Bertl maybe, let's see ... 1170447518 M * juggo ok I think I got it 1170447527 M * ebiederm Bertl: I'll probably have to look, one problem is that things that work fine in practice can have problems when the whole world is looking for security exploits. 1170447528 M * juggo -c xid pid 1170447565 M * Bertl ebiederm: well, having capabilities per space is nice, but opens some security holes by default, at least if you do not mask them properly :) 1170447596 M * ebiederm Bertl: Definitely. I'm haven't truly gotten farther than that there is a problem. 1170447606 M * Bertl juggo: depends on the signal, this will send 15 IIRC 1170447620 M * DotHack this is what i got with the kernel debian supllies: BUG: unable to handle kernel NULL pointer dereference at virtual address 00000000 1170447631 M * DotHack when killing a vserver 1170447636 M * Bertl DotHack: exactly, that's why I suggested to upgrade 1170447653 M * DotHack but only when the vserver is running etch and apache and mysql 1170447659 M * DotHack really weird 1170447660 M * Bertl about the vmlinuz-2.6.18.5-vs2.1.1.3-p3-squash-drbd-256ip-ipv6 I do not know anything ... 1170447674 M * daniel_hozac looks like an older derjohn kernel :) 1170447676 M * DotHack its from linux-vserver.org 1170447682 M * DotHack yeah derjohn 1170447684 M * daniel_hozac i can assure you it is not. 1170447695 M * daniel_hozac linux-vserver.org does not have _any_ binary kernels. 1170447709 M * Bertl (at least I hope so :) 1170447718 M * daniel_hozac it might link to quite a few, but there are none actually there. 1170447742 M * DotHack haha yeah there was a link 1170447787 M * Bertl you might want to try out the 2.2.0-rc10 (to be uploaded in a few minutes) though, should work perfectly fine for your setup 1170447814 M * derjohn i am happy that my stuff is used. :) I can confirm that kernel to be working on my machines 1170447843 M * daniel_hozac might want to jump on the 2.6.19 bandwagon ;) 1170447882 M * DotHack derjohn thanx! 1170447897 M * DotHack i dont like compiling kernels because i always forget stuff 1170447952 M * Bertl like what? 1170447979 M * daniel_hozac if you don't mind waiting, you could always do a kitchen-sink build. 1170447989 M * DotHack boh?? 1170448020 M * DotHack derjohn is this kernel ok for SMP? 1170448068 T * Bertl http://linux-vserver.org/ | latest stable 2.0.2.1, 2.0.3-rc1, 2.2.0-rc10, devel 2.1.1.7.1, 2.3.0.8, stable+grsec 2.0.2.1, 2.2.0-rc8, devel+grsec 2.1.1 | util-vserver-0.30.212 | 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 ;) 1170448074 M * Bertl *2.2.0-rc10 1170448075 M * derjohn DotHack, yes, I uses te same dotcondig as the org kernels 1170448089 M * derjohn hm, but limited to max 4 cpu ... (dual cpu dual core) 1170448093 A * derjohn away now 1170448108 M * derjohn must go, grrl just came here bye 1170448114 M * DotHack bye 1170448115 M * Bertl cya 1170448120 M * DotHack those girls 1170448126 Q * Piet Ping timeout: 480 seconds 1170448140 M * juggo i think for production systems people will be reluctant to always be installing the newest, relatively untested kernel 1170448161 M * Wonka hf, derjohn 1170448170 M * juggo i'd much rather see the problems fixed in whatever kernel is going to be used in debian 4.0 1170448172 M * daniel_hozac so it's better to go with the ones where you know what bugs to expect? :) 1170448185 M * Bertl juggo: which is somewhat odd, as you know that the older kernels have known bugs :) 1170448209 M * juggo right, known bugs are better than unknown bugs 1170448233 M * Bertl juggo: if you say so :) 1170448238 M * daniel_hozac so what if rebooting guests oopses the kernel and makes the context unusable, or what? 1170448246 M * DotHack i always liked the way debian does his thing with unstable,testing and stable but i really hope they dont screw up the new stable with these pesky bugs 1170448279 M * Bertl DotHack: according to juggo that is no problem, as long as you know what bugs are there :) 1170448283 M * juggo then that bug is critical, and needs to be patched in the stable kernel 1170448292 M * Wonka known bugs are patched out 1170448302 M * Wonka but no newer stuff is patched in 1170448312 M * Wonka that's what stable is for 1170448317 M * Bertl Wonka: so what about 2.6.18-3 ? 1170448319 M * DotHack and this is a known bug? 1170448333 M * juggo i'm just saying I've been pushing vservers in my work, but no one I work with is going to accept "oh that bug will go away if we just move to the kernel that was just released" 1170448337 M * Wonka Bertl: i don't know? 1170448339 M * DotHack i dint find it anywhere 1170448359 M * Bertl Wonka: I mean, it's a critical, known bug there, I don't see that it got replaced? 1170448373 M * juggo for production business systems I don't want to always be forced to run the latest kernel just to keep my system from crashing 1170448379 M * Wonka Bertl: was there a report filed? with a patch? 1170448392 M * daniel_hozac juggo: you expect us to retroactively fix bugs? 1170448393 M * Bertl Wonka: I'd say at least 10 should be there 1170448398 M * Wonka mh 1170448422 M * Bertl unless folks who said they did/will report it, forgot to do so 1170448424 M * juggo I would hope someone would patch critical bugs in a kernel that's going to be widely used 1170448457 M * juggo to a certain extent it depends on who your target audience is 1170448486 M * juggo if you're not interested in the business market then that's fine, I'll just start migrating things to xen ;) 1170448535 M * daniel_hozac juggo: unfortunately our time machine is in the shop... so we can't really fix bugs before we know they exist. 1170448548 M * Bertl juggo: who is 'you' and what do you consider 'business market'? 1170448551 M * juggo that's not what I've said at all 1170448558 M * daniel_hozac juggo: sure it is. 1170448571 M * juggo what I'm saying is that people are reporting crashes under 2.6.18-3 in etch 1170448576 M * juggo I've experienced them myself 1170448583 M * Bertl yes, and debian is not fixing them 1170448598 M * daniel_hozac we released the fixed version the 10th of december. 1170448598 M * Bertl what should we do about that? 1170448618 M * juggo what you have to do will all debian developers, keep bugging them 1170448627 M * juggo unless you can meet them in person, then buy them a beer 1170448632 M * daniel_hozac well, that's your job, no? 1170448648 M * juggo it should be all of our jobs 1170448649 M * daniel_hozac i'm not experiencing the problem. 1170448654 M * Bertl we cannot reasonably go and bug 20 distro folks every day 1170448673 M * juggo ok, only a suggestion that I see as being in your best interest 1170448675 M * Bertl just because they are _not_ doing their job properly 1170448683 Q * shedi Quit: Leaving 1170448688 M * Bertl juggo: what do you know about my interrest :) 1170448695 M * daniel_hozac juggo: how could we possibly influence Debian? 1170448725 M * juggo if those bugs aren't fixed, and people encounter them, they are not going to come away with a favorable impression of vservers 1170448748 M * Bertl juggo: I mainly want to a) have a working solution for me, and b) provide one for others to use ... 1170448753 M * juggo likewise, as I said, my business is launching virtual server hosting, and I can guarantee we are going to stick with debian 1170448775 M * Bertl juggo: the working solution is there, rock solid, take it or leave it :) 1170448781 M * juggo but whatever virtual solution we choose has to be stable under debian 1170448781 M * daniel_hozac juggo: so you should befriend yourself with the Debian maintainers, and make sure they're doing their job right, no? 1170448796 M * daniel_hozac really, we can't help what the Debian maintainers do and don't do. 1170448800 M * juggo well, you could influence the debian people that are in this channel, there are a few of them 1170448823 M * Bertl by kicking them every now and then? or what? 1170448833 M * juggo but it's true, if you just want something for yourself that may or may not work for other people that is your choice 1170448840 M * daniel_hozac doesn't help when the Debian kernel team has internal disagreements, does it? 1170448856 M * daniel_hozac seriously, you're talking to the wrong people. 1170448870 M * Bertl juggo: it works for us, and for you too, you just have to make the right decisions ... 1170448887 M * daniel_hozac you could avoid the whole Debian problem by simply compiling stuff yourself... 1170448902 M * Bertl and get a much more performant kernel, with less issues 1170448912 M * juggo and far less tested 1170448932 M * Bertl because mainline kernels are untested? comeon? 1170448934 M * juggo the reason to use the debian kernel is because there will be lots of machines using the same kernel 1170448944 M * daniel_hozac you really think the Debian vserver kernel has more users than the patches? 1170448946 M * DotHack juggo: yeah with the same bugs 1170448989 M * juggo it has the potential to have far mroe users 1170449003 M * juggo a number of people are running the vserver kernel on laptops to give the a dev environment 1170449017 M * Bertl as has redhat and mandriva and gentoo and slackware ... 1170449022 M * juggo I think that's a good application for vservers 1170449039 M * Bertl juggo: yes, it works perfectly fine for that 1170449080 M * juggo well, if it were me, spending a lot of time on a project, I'd want it to work well in all the major distros, 1170449109 M * Bertl juggo: but let me make the bottom line in this discussion, do something useful, complain to the Debian folks ... maybe when the number of bug reports reaches a critical limit, they will take some action ... 1170449134 M * Bertl until then, don't use the broken kernels or live with the bugs :) 1170449201 M * DotHack crappy debian folk 1170449215 A * mugwump is talking about containers & vserver in 15min 1170449226 M * daniel_hozac juggo: we really can't do any more than we already do. 1170449228 M * juggo fair enough, I'll check out the debian xen kernel to see if it's broken 1170449238 P * juggo 1170449246 M * ebiederm Well the other possibility is that non-one has submitted an updated debian package to the debian folk. 1170449253 M * Bertl DotHack: I think it is more an internal issue, than a real people problem, most debian folks are quite okay ... 1170449273 M * Bertl ebiederm: there is a 2.6.18-4 kernel from debian folks, for debian, with the fixes ... 1170449295 M * Bertl ebiederm: it just doesn't get in because of 'unknown internal issues' 1170449305 J * Aiken ~james@ppp220-70.lns2.bne1.internode.on.net 1170449311 M * Bertl morning Aiken! 1170449313 M * ebiederm Bertl: That is weird. 1170449318 M * DotHack crappy internal issues 1170449322 M * Aiken hello 1170449326 M * Bertl DotHack: indeed! 1170449341 M * daniel_hozac i would assume the problem was an ABI-bump so close to the release. 1170449353 M * daniel_hozac (but that hasn't happened yet, so...) 1170449392 M * Bertl don't know, don't care ... 1170449400 J * gerrit ~gerrit@bi01p1.co.us.ibm.com 1170449404 M * ebiederm This reminds me that stability doesn't mean bug-free just that things have been around long enough the bugs are known :) 1170449422 M * daniel_hozac indeed. 1170449487 M * Bertl one strange assumption (which I really cannot understand) is that kernels get better when they get older :) 1170449506 M * Bertl it's not wine, damnit! :) 1170449516 M * daniel_hozac yeah, i've never understood that concept. 1170449542 M * daniel_hozac i'd like to think we fix more bugs than we create... 1170449557 M * DotHack wine as in wine is not an emulator :p 1170449568 M * Bertl I'm not sure this will be true for 2.6.20 though :) 1170449598 M * ebiederm What concerns you about 2.6.20 ? 1170449609 M * Bertl too many changes in subtle places 1170449612 M * DotHack argh this is the 5 time my internet breaks 1170449627 M * DotHack luckiliy i have screen hehe 1170449643 M * Bertl ebiederm: I assume we will not be able to find certain bugs for several months 1170449658 M * Bertl s/find/discover/ 1170449683 M * ebiederm Bertl: Which subtle places are you thinking of? 1170449698 M * Bertl all the thread changes, and pid preparations 1170449734 M * ebiederm The pid preparations that were problematic were reverted, so I think we are ok on that front. 1170449735 M * Bertl already 2.6.18->2.6.19 did break certain assumptions in somewhat unexpected ways ... 1170449738 M * ebiederm Thread changes? 1170449747 M * Bertl kthread work and such 1170449756 M * ebiederm Bertl: Ok. 1170449795 M * ebiederm Bertl: Yes the kthread work could have a few negative consequences. 1170449819 M * ebiederm I think it is all less likely to be an issue then when I complete redid the irq handling sever kernels back on x86_64 :) 1170449827 M * ebiederm Right when Ingo was doing the genapic work. 1170449870 M * Bertl I'm not saying it will explode in your face ... immediately, but there will be breakage in otherwise stable places ... 1170449972 M * ebiederm Bertl: Probably, the danger of touching code that hasn't been touched in a long time. Luckily most of the breakage will be isolated to silly little corners where no-one goes anyway. 1170450014 M * Bertl well, that is 'luckily' for the folks out there, but doesn't affect the number of actual bugs, no? 1170450079 M * ebiederm Actually I think the kthread work is likely to be a wash in the number of bugs. Some of that code was questionable to start with. 1170450148 M * Wonka Bertl: http://packages.debian.org/changelogs/pool/main/l/linux-2.6/linux-2.6_2.6.18.dfsg.1-9/changelog 1170450149 M * ebiederm Anyway we will see. 1170450190 M * Bertl Wonka: so that means it should be resolved soon? 1170450199 M * Wonka Bertl: at least, there is a -4 1170450220 M * Bertl Wonka: that's what I said, no? but it is not in debian yet 1170450223 M * daniel_hozac Wonka: yes, but it's only in sid. 1170450266 M * Wonka daniel_hozac: http://packages.qa.debian.org/l/linux-2.6.html says "too young, only 8 of 10 days old" 1170450282 M * Wonka daniel_hozac: also, "Not touching package, as requested by freeze (contact debian-release if update is needed)" 1170450284 M * daniel_hozac the packages have been available from other sources for far longer. 1170450297 M * daniel_hozac (kernel-archive.buildserver.net or what was it?) 1170450301 M * Bertl yes, IIRC, mid of december 1170450301 M * ebiederm Bertl: Probably the scary bug is the ipc namespace exiting one that Kirill seems to have spotted. 1170450337 M * Bertl ebiederm: haven't got to that yet (I think?) 1170450342 M * Bertl ebiederm: details? 1170450376 M * Bertl the ipc stuff seems to work reasonably fine with 2.6.19.2 here 1170450388 M * ebiederm Bertl: Something he posted on the containers list. I haven't had a chance to look into it yet. But basically he believes you can form a cycle that defeats the reference counting. 1170450432 M * ebiederm There weren't enough details when he posted to make independent verification easy. 1170450461 M * Bertl well, why doesn't that surprise me ... 1170450465 M * ebiederm Something about a shm_segment pointing at a file pointing at a shm_segment or something like that. 1170450498 M * ebiederm Anyway the message is only a day or two old in the containers list. 1170450507 M * Bertl okay, will check it, tx 1170450512 M * ebiederm Bertl: Kirillil was stumped for a fix and asking for help. 1170450542 M * Bertl kir: ping? 1170450551 M * Bertl (maybe he is around :) 1170450674 M * daniel_hozac http://lists.osdl.org/pipermail/containers/2007-January/001197.html is the mail, FWIW. 1170450692 M * Bertl great! tx! 1170450750 M * Bertl ebiederm: we also use two counters, one for tasks and one for the reference 1170450761 M * Bertl ebiederm: at least for the process and network contexts 1170450787 M * Bertl this also helps with controlling context/space visibility and reuse 1170450804 M * ebiederm Bertl: There are certainly times that make sense. 1170450815 M * Bertl i.e. you can hide away a context, even when it is kept around for example for sockets (as reference) 1170450877 J * xp_prg ~xp_prg@ftp.microvu.com 1170450884 M * Bertl welcome xp_prg! 1170450894 M * xp_prg hi all I am new to vserver stuff, can I get newbie help here? 1170450900 M * Bertl sure 1170450932 M * xp_prg well I have a box with debian on it and vserver patches compiled into the kernel, I now want to create a centos 4.4 guest os vserver I am not clear on the best way to do this 1170450940 M * xp_prg I am familiar with vmware 1170450941 M * ebiederm Thinking about it. Two counters could work in this instance but not a process and an other counter but more of a counter to keep this thing alive, and a count to make certain I don't have leaks kind of counter. 1170450953 M * xp_prg in vmware I can just download a centos 4.4 vmware appliance and run it 1170450962 M * Bertl xp_prg: first, forget most you know about VMware 1170450964 M * xp_prg does such a thing exist for vserver? 1170450966 M * DotHack finally files the bug to debian 1170450966 M * xp_prg ok 1170450972 M * DotHack crappy internet connection 1170450994 M * Bertl xp_prg: then, if you have a centos system (installed or virtual from vmware) 1170451005 M * xp_prg virtual from vmware 1170451006 M * Bertl xp_prg: you can take this and extract the files to a directory 1170451029 M * Bertl after a minor cleanup (removal of most of the /dev entries and some runlevel scripts) 1170451045 M * Bertl you can start this with a proper Linux-VServer guest configuration 1170451050 M * daniel_hozac you'll likely end up with a cleaner guest if you do apt-get install yum; vserver centos build -m yum --context ... --hostname ... --interface ... -- -d centos4 1170451080 M * xp_prg daniel_hozac wow that sounds easy 1170451084 M * Bertl that is another way to go, which will require yum to be installed 1170451105 M * Bertl it has the advantage that you get a newly installed guest from the internet 1170451116 M * xp_prg let me try that 1170451130 M * Bertl xp_prg: you need to replace the ... with values 1170451146 M * xp_prg right 1170451150 M * Bertl xp_prg: i.e. select a context id (2-49151) a host/domain name 1170451168 M * xp_prg ok 1170451169 M * Bertl xp_prg: and specify an interface with ip, e.g. --interface eth0:10.0.0.1/24 1170451180 M * xp_prg thanks will do, I appreciate this help :> 1170451187 M * Bertl you're welcome! 1170451200 M * Bertl most important: don't think VMware, think Linux :) 1170451202 A * xp_prg wonders why there can't be like vserver appliances you can just download and run is that not possible at all? 1170451212 M * xp_prg well I ran vmware on linux at least :) 1170451230 M * Bertl xp_prg: yes, but would you like to download my handcrafted keyboard logger and DDoS botnet? 1170451265 M * Bertl xp_prg: if so, please just drop me a note and I provide a binary image for you :) 1170451295 M * xp_prg Bertl ok, can you help me to understand how you create this "binary image" that I can download? 1170451314 M * daniel_hozac just create a tarball or similar. 1170451321 M * Bertl sure, I would just do what daniel_hozac suggested, then copy a few malicious files into the tree 1170451332 M * Bertl after that, I'd make a tar for you to extract :) 1170451349 M * daniel_hozac then you can extract it with vserver blah build -m template ... -- -t /path/to/tarball 1170451362 M * xp_prg cool 1170451364 M * Bertl exactly, ... is again some options 1170451380 M * xp_prg is there like a central repository for these kind of vserver appliances? 1170451395 M * daniel_hozac no, the idea is that you can build them yourself. 1170451404 M * daniel_hozac to suit _your_ needs. 1170451413 M * daniel_hozac rather than, e.g., my needs :) 1170451491 M * Bertl xp_prg: it's not really hard to do, just give it a try :) 1170451518 M * xp_prg vserver centos build -m yum --context 8000 --hostname pricelist --interface pricelist=eth0:10.0.0.1/24 -d centos 1170451523 M * xp_prg this is not working ;( 1170451532 M * xp_prg yum is intalled however 1170451532 M * daniel_hozac you're missing a -- before -d 1170451545 M * daniel_hozac and a 4 after centos. 1170451546 M * xp_prg ERROR: Can not find configuration for the distribution 'centos'; please read http://linux-vserver.org/HowToRegisterNewDistributions for information how to add support for your own distribution. 1170451564 M * daniel_hozac yep, as i said, you need -d centos4. 1170451570 M * Bertl copy & paste error :) 1170451580 M * xp_prg mount: mount point /etc/rpm does not exist 1170451597 M * daniel_hozac and you did install yum? did it not install rpm with it? 1170451613 M * xp_prg I don't know, I did install yum though 1170451615 M * Bertl xp_prg: debian? if so, try apt-get install yum 1170451634 M * daniel_hozac (i only have a Fedora hosts, so i'm not exactly sure what's needed on other distributions) 1170451668 M * xp_prg Install 7 Package(s) Update 0 Package(s) Remove 0 Package(s) Total download size: 21 M 1170451674 M * xp_prg I guess it is woring 1170451684 M * daniel_hozac looks like. 1170451744 M * xp_prg yes it is getting the packages 1170451755 M * xp_prg which rpm /usr/bin/rpm 1170451759 M * xp_prg rpm is in my path 1170451778 M * xp_prg how long do you think it will take daniel_hozac? 1170451795 M * daniel_hozac depends on your connection to the mirror. 1170451811 M * daniel_hozac the other stuff shouldn't take more than 5 minutes or so. 1170451825 M * xp_prg do I need to like configure other things that I can do while this is working? 1170451832 J * Administrator ~chatzilla@node116.hs.phippu.ch 1170451848 M * daniel_hozac such as? that should be it. 1170451861 M * Administrator g 1170451876 M * xp_prg well I heard I had to like make a kind of /etc/hosts entry or something for the ip to work 1170451892 N * Administrator weeze 1170451923 M * xp_prg Cannot find device "eth0" SIOCGIFFLAGS: No such device 1170451929 M * xp_prg it just failed with this error 1170451959 M * daniel_hozac do you have an eth0? 1170451966 M * DotHack what is the difference between xen and vserver? 1170451973 M * xp_prg in my host I do 1170452007 M * daniel_hozac and at what point did it output that? 1170452021 M * daniel_hozac DotHack: para/full virtualization vs. isolation and virtualization where required. 1170452030 M * xp_prg can I flood in here? 1170452034 M * daniel_hozac no. 1170452040 M * daniel_hozac use paste.linux-vserver.org. 1170452052 M * weeze hi when i try to set routes in a vserver i get this error message "SIOCDELRT: Operation not permitted" any ideas ? 1170452072 M * weeze where do i set the default gw of a vserver ? 1170452072 M * daniel_hozac weeze: that's because networking happens on the host. 1170452077 M * FaUl on the host 1170452101 M * weeze ok, host has a default gw..but i cannot recach beyond that gw from the vserver 1170452111 M * daniel_hozac why not? 1170452112 M * xp_prg http://paste.linux-vserver.org/1019 1170452137 M * weeze s/reach/recach 1170452177 M * daniel_hozac xp_prg: looks like the guest should've built successfully. as to why you get that error, i have no idea. 1170452181 M * weeze so where do i set the default gw of a vserver on the host ? 1170452193 M * daniel_hozac weeze: the guest will use the host's default gateway. 1170452195 M * DotHack route 1170452200 M * weeze hmmm 1170452209 M * daniel_hozac weeze: just like it will use all of the other routes. 1170452310 M * xp_prg daniel_hozac oh wait there is no eth0 1170452328 M * daniel_hozac there you go then... 1170452334 M * xp_prg there is an eth1 and an eth2 1170452386 M * Bertl you can easily change that in the config now 1170452401 M * Bertl go to /etc/vservers//interfaces/0 1170452409 M * Bertl and change the 'dev' entry accordingly 1170452419 J * baldy 1000@83.223.66.130 1170452420 M * baldy re 1170452425 M * xp_prg vserver-topdirectory '/etc/vservers/.defaults/vdirbase/centos4/etc' and/or configuration at '/etc/vservers/centos4' exist already; please try to use '--force', or remove them manually. 1170452426 M * Bertl wb baldy! 1170452439 M * xp_prg I try: vserver --force centos4 build -m yum --context 8000 --hostname pricelist --interface pricelist=eth1:10.0.0.1/24 -- -d centos4 1170452440 M * Bertl xp_prg: no need to rebuild the guest 1170452451 M * baldy Bertl: works fine with xid ;) 1170452457 M * xp_prg Bert1 what should I do now then? 1170452467 M * Bertl baldy: great! 1170452475 M * Bertl xp_prg: did you already use the --force? 1170452490 M * baldy when dm8tbr is here we will test IPv6 tun/tao devices ;) 1170452495 M * xp_prg /usr/sbin/vserver: unrecognised option `--force' 1170452503 M * xp_prg so I guess not since that didn't work 1170452506 M * Bertl okay, so you are lucky :) 1170452514 M * xp_prg so what do I do now then? 1170452523 M * Bertl well, it would work with --force too, but I would start downloading your guest again 1170452530 M * dm8tbr baldy: at your service! 1170452535 M * Bertl xp_prg: do what I suggested, in the config 1170452545 M * Bertl 22:40 < Bertl> go to /etc/vservers//interfaces/0 1170452545 M * Bertl 22:40 < Bertl> and change the 'dev' entry accordingly 1170452569 M * baldy dm8tbr: i dont have native v6 in this location atm.. 1170452576 M * baldy i must setup the router first hehe 1170452611 M * dm8tbr baldy: ok, so are we going to try tun/tap first? 1170452654 M * xp_prg ok I did that Bertl now what do I do? vserver centos4 start? 1170452655 M * baldy yep 1170452663 M * baldy tell me when ;) 1170452708 M * Bertl xp_prg: yep, I hope you stopped it before ... 1170452727 M * Bertl if not, do the stop now, it will give an error, but that should be harmless 1170452822 M * xp_prg Bert1 ok it started 1170452826 M * xp_prg do I now enter it? 1170452835 M * Bertl if you like to, yes :) 1170452839 M * xp_prg oh cool I am in it now 1170452871 M * xp_prg Bertl I am not clear on how to continue the downloading of the full centos system, can you help me, I beleive it is not complete right? 1170452885 M * daniel_hozac xp_prg: complete in what way? 1170452900 M * xp_prg downloading and getting centos and all the packages etc... 1170452903 M * Bertl depends on what you want to do/get, centos has something about 4GB packages or so= 1170452917 M * xp_prg you mean I have a fully functioning centos4 right now? 1170452919 M * Bertl that will probably make up 8GB of disk space or more 1170452932 M * Bertl if you need that stuff, of course :) 1170452951 M * daniel_hozac xp_prg: the stuff that's installed works, no? 1170452951 M * Bertl xp_prg: typically you just install the packages you want to use 1170452974 M * xp_prg bash: yum: command not found 1170452977 M * Bertl xp_prg: a minimal, but working centos is already installed 1170452989 M * daniel_hozac xp_prg: you need to do that from the host. 1170452992 M * Bertl the package management is external for now 1170452997 M * daniel_hozac xp_prg: until you've internalized package management, at least. 1170453009 M * xp_prg Bertl my host is debian 1170453032 M * Bertl you can switch that inside the guest, if you like, but make sure to install yum/rpm there 1170453048 M * Bertl xp_prg: nevertheless, external management works on debian too 1170453061 M * daniel_hozac vyum -- install yum; vserver pkgmgmt internalize, for instance. 1170453083 M * Bertl that will move the package management into the guest, after installing yum there :) 1170453103 M * Bertl of course, you can 'just' use vyum to install packages from outside 1170453126 M * daniel_hozac which keeps the guest harddrive requirements down. 1170453283 M * Bertl xp_prg: the next issue you will encounter (according to my crystall ball) is that you try to ssh to the guest, but it doesn't work and/or you reach the host :) 1170453317 M * Bertl xp_prg: ah, no, now I see it clearly, you cannot reach the internet :) 1170453352 P * suka Leaving 1170453420 M * xp_prg Bertl I did vserver centos4 enter 1170453426 M * xp_prg there is no which 1170453438 M * daniel_hozac nope. 1170453442 M * xp_prg ls exists 1170453449 M * daniel_hozac you'd need to install it first. 1170453468 M * xp_prg how do I get access to yum again? 1170453473 M * daniel_hozac or you could just use type -p, which just uses bash. 1170453475 M * xp_prg inside the vserver? 1170453487 M * daniel_hozac did you run the command above? 1170453492 M * Bertl xp_prg: do you have an irc backlog? 1170453516 M * xp_prg yes 1170453519 M * weeze hello routing works fine now...hosts iptables and the roters nating were the issue 1170453524 M * weeze tnx 4 help 1170453526 M * xp_prg so sounds like vyum is what I want to use 1170453543 M * Bertl either that, or internalize the package management, yes 1170453580 M * xp_prg vyum centos4 -- install yum 1170453596 M * xp_prg that will install yum into my centos4 vserver yes? 1170453607 M * daniel_hozac yes. 1170453622 M * xp_prg thankyou for your patience I know newbies can be irritating :) 1170453638 M * Bertl np, that's fine ... 1170453766 M * xp_prg http://paste.linux-vserver.org/1020 1170453770 M * xp_prg that command failed 1170453808 M * daniel_hozac that sounds like your guest doesn't know how to resolve names. 1170453830 M * daniel_hozac did you assign the guest a usable IP address? 1170453957 M * xp_prg daniel_hozac now I am confused I thought I assigned the ip address with: vserver centos4 build -m yum --context 8000 --hostname pricelist --interface pricelist=eth1:10.0.0.1/24 -- -d centos4 1170453975 M * daniel_hozac but can you access the internet from 10.0.0.1? 1170453988 M * daniel_hozac i.e. does ping -I 10.0.0.1 google.com work on the host? 1170454009 M * xp_prg ping -I 10.0.0.1 google.com PING google.com (64.233.167.99) from 10.0.0.1 : 56(84) bytes of data. 1170454016 M * xp_prg no ping 1170454018 M * daniel_hozac as Bertl's crystal ball predicted 10 minutes ago :) 1170454022 M * xp_prg heh 1170454117 M * xp_prg how can I fix that? 1170454130 M * daniel_hozac do you have a public IP address assigned to eth1 right now? 1170454157 M * xp_prg no 1170454161 M * xp_prg eth2 has one 1170454187 M * daniel_hozac so eth1 has an address on the 10.0.0.0 network? 1170454193 M * xp_prg wait now I am confused, how can I tell? 1170454203 M * daniel_hozac ip a 1170454206 M * xp_prg eth1 Link encap:Ethernet HWaddr 00:15:F2:CA:CE:F7 inet addr:172.16.1.2 Bcast:172.16.1.255 Mask:255.255.255.0 inet6 addr: fe80::215:f2ff:feca:cef7/64 Scope:Link 1170454239 M * xp_prg I logged into this machine with ssh -X -l root 172.16.1.2 1170454272 M * daniel_hozac so where do you have the 10.0.0.0 network? 1170454281 M * daniel_hozac or don't you have one, and wish to use it for your guests? 1170454299 M * xp_prg I don't have one, and I wish to use it for my guests 1170454310 Q * gerrit Ping timeout: 480 seconds 1170454336 M * daniel_hozac so, you'll have to configure your firewall to NAT that network as well. 1170454349 M * daniel_hozac (i assume this box is already performing NAT, right?) 1170454359 M * xp_prg oh man no 1170454360 Q * dhansen Ping timeout: 480 seconds 1170454361 M * Bertl probably outside the host 1170454373 M * xp_prg can I just give it like 172.16.1.5 or something? 1170454379 M * xp_prg whatever is easiest for now please 1170454384 M * Bertl yes, but will that become public? 1170454399 M * Bertl i.e. does your host allow to use the 172.16.1.5 ip too? 1170454400 M * xp_prg no, this is behind a router doing nat 1170454411 M * xp_prg dhcp is turned off in the router 1170454418 M * daniel_hozac well then, just use some more of those addresses. 1170454422 M * Bertl would the router NAT 172.16.1.5 ? 1170454427 J * duckx ~Duck@tox.dyndns.org 1170454439 M * xp_prg Bert1 I don't know, how do I tell? 1170454451 M * xp_prg it is nat 172.16.1.2 and 172.16.1.3 1170454464 M * Bertl so for those two it already works? 1170454468 M * xp_prg yes 1170454476 M * Bertl and both are used by the host? 1170454489 M * xp_prg no, those are other computers connected to the router 1170454504 M * Bertl well, you don't want to use their IP :) 1170454505 M * xp_prg the host's ip address is 172.16.1.2 1170454520 M * xp_prg right so I was thinking of giving the guest os 172.16.1.5 does that sound good? 1170454523 M * Bertl who does control the router? 1170454526 M * xp_prg me 1170454540 M * Bertl then add a NAT rule/entry for 172.16.1.5 on the router 1170454554 M * Bertl and you will be good, after stopping the guest, and chaning the ip to that 1170454564 M * xp_prg wow never done that before, is it not automatic? 1170454576 M * Bertl xp_prg: I don't know your router :) 1170454581 M * xp_prg it is a netgear 1170454593 M * Bertl nice :) 1170454597 M * xp_prg oh well, how can I see if it could work? would do ping -I 172.16.1.5? 1170454616 Q * dmax Ping timeout: 480 seconds 1170454620 M * Bertl yes, if you try with ping -I 172.16.1.5 www.google.com on the host 1170454631 M * Bertl that is a good check, given that the 172.16.1.5 is assigned somewhere 1170454633 M * daniel_hozac you'd have to assign the address first. 1170454638 M * Bertl (check with ip addr ls) 1170454642 M * daniel_hozac heh 1170454669 M * xp_prg bind: Cannot assign requested address 1170454680 M * daniel_hozac did you assign the address? 1170454687 M * xp_prg I don't think so, how do I do that? 1170454721 M * daniel_hozac ip addr add
/ dev ethX 1170454781 M * xp_prg ip addr add 172.16.1.5/255.255.255.0 dev eth1 Error: an inet prefix is expected rather than "172.16.1.5/255.255.255.0" 1170454794 M * daniel_hozac guess it requires prefix, then. 1170454805 M * xp_prg what is the prefix? 1170454815 M * daniel_hozac 24 1170454844 M * xp_prg ok that worked! 1170454854 M * xp_prg it pings with ping -I 172.16.1.5 www.google.com 1170454864 M * xp_prg I guess I must now delete that entry? 1170454881 A * xp_prg doesn't really understand what he just did 1170454907 M * Bertl well, you added a new ip to the interface 1170454919 M * ebiederm Hmm. mugwump must have meant he is giving a presentation right now. 1170454924 M * Bertl you can remove it now with ip addr del 172.16.1.5/24 dev eth0 1170454933 M * Bertl ebiederm: yes 1170455061 M * xp_prg the vyum command is not working :) 1170455068 M * xp_prg I mean now working 1170455215 M * xp_prg ok I have yum now inside my centos4 1170455225 M * xp_prg now I guess I need to configure yum 1170455414 M * Bertl IIRC, it should get the proper config 1170455428 M * xp_prg yum install ssh Setting up Install Process Setting up repositories not using ftp, http[s], or file for repos, skipping - Null is not a valid release or hasnt been released yet Cannot find a valid baseurl for repo: update Error: Cannot find a valid baseurl for repo: update 1170455444 M * xp_prg any idea how to fix that? 1170455466 M * Bertl did you internalize the management yet? 1170455535 M * xp_prg I don't think so, how can I tell? 1170455567 M * Bertl well, you should know what you executed so far, if not, 'history' might help 1170455683 M * xp_prg I don't think I internalized the management yet, I am scrolling but having a hard time finding that command, can you please repeat it? 1170455702 M * xp_prg I can now ping from within my vserver :) 1170455724 M * Bertl vserver pkgmgmt internalize 1170455839 M * xp_prg http://paste.linux-vserver.org/1021 1170455847 M * xp_prg yum install ssh attempts but fails 1170455895 M * Bertl maybe daniel_hozac has some ideas, looks like a centos issue though 1170455912 M * daniel_hozac even after internalizing? 1170455918 M * xp_prg yes 1170455931 M * daniel_hozac rm -f /var/lib/rpm/__db*; rpm --rebuilddb and retry. 1170455963 M * xp_prg now it is working looks like :) 1170456047 P * stefani I'm Parting (the water) 1170456073 M * xp_prg wahoo it is working, you guys are soooooo smart thanks 1170456084 M * Bertl you're welcome! 1170456100 M * xp_prg do you know what yum package has which in it? 1170456115 M * Bertl IIRC, there is a search option for yum? 1170456133 M * xp_prg yes I am using it 1170456152 M * Bertl well, it should find what you are looking for then :) 1170456202 M * xp_prg how did you guys get sooo good? 1170456242 M * Bertl well, you lear one or two things if you actually work with Linux :) 1170456259 M * xp_prg beleive it or not I have been working with linux for years but don't know the stuff you told me 1170456318 M * Bertl well, I guess some of it is relatively Linux-VServer specific 1170456492 M * Bertl daniel_hozac: ever seen an error in syscall32_nopage on x86_64? 1170456507 M * daniel_hozac not that i can remember. 1170456508 M * Bertl I have a 64bit kernel here, with 32bit userspace 1170456518 M * Bertl everything is 32bit in userspace for now 1170456532 M * Bertl and it seems that hotplug bails out with that every time 1170456541 M * daniel_hozac hmm, strange. 1170456557 M * Bertl [ 131.587662] Pid: 191:#0, comm: hotplug Not tainted 2.6.19.2-vs220_64 #2 1170456567 M * Bertl [ 131.778087] Pid: 193:#0, comm: hotplug Not tainted 2.6.19.2-vs220_64 #2 1170456602 M * Bertl except for that, it is running fine though :) 1170456881 M * Bertl ah, maybe 2/2 doesn't work with x86_64 kernels and ia32? 1170456893 J * FireEgl ~FireEgl@2001:5c0:84dc:1:211:9ff:feca:b042 1170456962 M * daniel_hozac 2/2 memory split? 1170456988 M * Bertl yep 1170457023 M * daniel_hozac i didn't even know x86_64 had that. 1170457047 M * Bertl hmm, right, I had it in the config, but it got auto removed, so that should be fine 1170457062 M * daniel_hozac could you take a look at /src/SYSCALL/test-syscall.c? am i doing something wrong, or are the syscalls not working right? 1170457068 M * waldi x86_64 have a fixed split 1170457090 M * waldi anything with bit 48 or so set is kernel 1170457099 M * waldi or maybe it was 47 1170457115 M * daniel_hozac (socketpair gets called with 0 for the fourth argument) 1170457121 M * Bertl 99.338568] Unable to handle kernel paging request at ffff82bc81000000 1170457129 M * Bertl [ 99.338572] [] syscall32_nopage+0x4c/0x64 1170457143 M * Bertl RIP: 0010:[] 1170457164 M * waldi the first address looks a little bit of the line 1170457175 M * waldi shifted 2 bytes 1170457196 M * Bertl page size issues? 1170457264 M * waldi dunno 1170457408 J * dmax ~semaj@81.193.58.121 1170458049 M * baldy Bertl: daniel_hozac are u there? 1170458060 M * baldy did u need the chbind6 tool that IPv6 work? 1170458074 M * baldy i tried to make following: 1170458075 M * daniel_hozac linux-vserver.org/IPv6 1170458078 M * baldy http://linux-vserver.org/IPv6 1170458091 M * baldy error is: 1170458092 M * baldy Broadcast can be set only for IPv4 addresses 1170458093 M * baldy Invalid IP number or host name '2a01:78:4::1' 1170458105 M * daniel_hozac and you _are_ using util-vserver 0.30.212+, right? 1170458125 M * baldy argh nope 1170458128 M * baldy fuck 1170458290 M * Bertl hmm? 1170458341 M * baldy openvcp works with 0.30.210 :( 1170458354 M * daniel_hozac that sounds like a bad move for everyone involved. 1170458369 Q * hallyn Quit: leaving 1170458418 J * dna_ ~naucki@240-229-dsl.kielnet.net 1170458541 Q * DotHack Quit: Chatzilla 0.9.77 [Firefox 2.0.0.1/2006120418] 1170458558 M * baldy i will check it know if it realy dont work 1170458584 J * mire ~mire@178-167-222-85.adsl.verat.net 1170458598 M * baldy s/know/now 1170458626 J * DavidS ~david@chello062178045213.16.11.tuwien.teleweb.at 1170458753 Q * dna Ping timeout: 480 seconds 1170458765 M * baldy mh is there a way to check which version of vserver-utils are running? 1170458769 M * baldy somethink lie vserver -V 1170458780 Q * ensc Ping timeout: 480 seconds 1170458809 Q * duckx Remote host closed the connection 1170458883 M * daniel_hozac like vserver --version? 1170458906 M * Bertl okay, off to bed ... have to get up early tomorrow ... 1170458914 M * Bertl have a good one everyone! cya! 1170458918 N * Bertl Bertl_zZ 1170458929 M * daniel_hozac good night! 1170458979 M * baldy Broadcast can be set only for IPv4 addresses 1170458979 M * baldy Invalid IP number or host name '2a01:78:4::1' 1170458986 M * baldy some prob as before 1170458988 M * baldy ohh 1170458992 M * baldy n8 guys 1170459032 M * daniel_hozac and you did upgrade? 1170459220 M * baldy yep 1170459230 M * daniel_hozac so vserver --version now shows 0.30.212? 1170459234 M * baldy now openvcp told "chbind: kernel does not provide network virtualization" 1170459250 M * daniel_hozac ?! what kernel are you using? 1170459267 M * baldy 2.6.19.2-vs2.2.0-rc9.ipv6 1170459287 Q * dna_ Quit: Verlassend 1170459290 M * daniel_hozac and how on earth did you configure that? 1170459308 M * baldy with the 210 version it runs 1170459310 M * baldy funny is 1170459319 M * baldy when i use the test.sh 1170459325 M * baldy everthink looks okey 1170459327 M * daniel_hozac probably because you enabled the legacy stuff in that build. 1170459346 M * daniel_hozac how are you installing the utils? 1170459375 M * baldy ./configure --prefix=/usr --sysconfdir=/etc --localstatedir=/var 1170459422 M * daniel_hozac and same command every time? 1170459428 M * baldy yep 1170459438 M * baldy http://www.openvcp.org/wiki/Install0.2/util-vserver 1170459443 M * baldy thats what openvcp said 1170459447 M * baldy 212 wont work 1170459494 M * daniel_hozac i guess you're screwed then. 1170459536 M * baldy i need a Webinterface hehe 1170459545 M * baldy and openvcp seems the best solution 1170459552 M * baldy or is there any other interface? 1170459693 M * daniel_hozac there's openvps (but that's even older, so probably even less likely to work). 1170459699 M * daniel_hozac and vsmon. 1170459710 M * daniel_hozac (http://vsmon.revolutionlinux.com/) 1170459715 M * daniel_hozac i haven't used any of them though. 1170459860 Q * DavidS Quit: Leaving. 1170459870 J * gerrit ~gerrit@bi01p1.co.us.ibm.com