1118188830 M * Bertl otherwise the processes start disk I/Os which they can not handle properly (within the short timeframe) which ultimately leads to 100% I/O waiting ... 1118188889 M * mugwump oh crap 1118188907 M * Bertl hmm? 1118188927 M * mugwump right, so I guess with hard limits it's actually desirable to make sure the fillrate is at least the time of an "average" IO 1118189017 M * Aiken doing a 'make -j 3' for a kernel build in the vserver. the host is holding 30 - 45% 1118189033 M * Aiken the host is that much idle 1118189070 M * mugwump wait, we're not deducting tokens from W state processes are we? 1118189087 Q * Doener_zZz Quit: Leaving 1118189310 M * Aiken not getting upto 49% idle, had a couple of firefox running on the host 1118189345 M * Aiken not sure if I was imaging it or not but the few firsfox seemed to have a higher cpu usage with the vserver running 1118189481 M * Aiken maybe imagination 1118189679 J * matti matti@linux.gentoo.pl 1118189858 M * Aiken I had not expected firefox to be hitting the cpu usage it was, take it out of the equation and the cpu usage is as expected 1118189946 Q * Zoiah Ping timeout: 480 seconds 1118190546 Q * Bertl Ping timeout: 480 seconds 1118191708 M * mugwump so, the main server is not limited to anything.. 1118192394 M * Aiken me? no limits on the host 1118192444 J * explasm__ explasm@p549FF503.dip.t-dialin.net 1118192504 J * Bertl ~herbert@janus.mc.tuwien.ac.at 1118192550 M * Bertl cool, it seems I was gone for almost an hour ... 1118192569 M * Bertl but now, suddenly my irc client caught up ... 1118192665 M * Bertl anyway, I'm off to bed for today ... back tomorrow ... 1118192676 M * Bertl have a nice whatever everyone! cya 1118192680 A * mugwump waves 1118192686 N * Bertl Bertl_zZ 1118192883 Q * eXplasm2 Ping timeout: 480 seconds 1118195588 M * Aiken if there a 1.2.10 patch for 2.4.31 kernel or just use the 2.4.30 patch? 1118203180 J * Aiken_ ~james@tooax8-222.dialup.optusnet.com.au 1118203506 Q * Aiken Ping timeout: 480 seconds 1118204206 J * monrad ~monrad@0x535b06c0.ronxx3.adsl-dhcp.tele.dk 1118208775 Q * Loki|muh Remote host closed the connection 1118208853 J * Loki|muh loki@satanix.de 1118209133 M * eyck 2.4.31.patch is very small, 1.2.10.1 applies without warning 1118209189 Q * Hollow Quit: /dev/zero 1118209460 J * Hollow ~Hollow@home.xnull.de 1118209536 Q * monrad Ping timeout: 480 seconds 1118209547 J * monrad ~monrad@0x535b06c0.ronxx3.adsl-dhcp.tele.dk 1118209565 M * Aiken_ where is 1.2.10.1? 1118209826 M * eyck it's a secret, and I have to run now :( 1118209832 M * eyck 'm almost late 1118211590 Q * monrad Quit: Leaving 1118211847 M * eyck Aiken_: http://vserver.13thfloor.at/Experimental/delta-2.4.30-vs1.2.10-vs1.2.10.1.diff 1118211981 M * Aiken_ thanks 1118212411 J * Doener ~doener@p5487446C.dip.t-dialin.net 1118212589 J * IceTi 125@mw635449.rz.ruhr-uni-bochum.de 1118214855 J * Aiken ~james@tooax6-205.dialup.optusnet.com.au 1118215163 Q * Aiken_ Ping timeout: 480 seconds 1118216753 Q * Aiken Ping timeout: 480 seconds 1118218303 J * Zoiah Zoiah@matryoshka.zoiah.net 1118218760 J * prae ~prae@ezoffice.mandriva.com 1118220655 J * monrad ~monrad@195.97.130.53 1118222861 J * rs ~rs@staff.lycos.fr 1118224379 J * jsambrook ~jsambrook@host-62-69-64-93.bsve.net 1118224382 P * jsambrook 1118227458 J * Doener` ~doener@p54874452.dip.t-dialin.net 1118227898 Q * Doener Ping timeout: 480 seconds 1118228049 Q * rs Quit: rs 1118230808 N * Bertl_zZ Bertl 1118230818 M * Bertl morning folks! 1118230920 M * mugwump hi Bertl 1118230931 M * Bertl hey mugwump! still up? 1118231397 M * IceTi hey 1118231429 M * Bertl hey Timo! LTNS! 1118231515 M * mugwump I'm still up, but on a slow link... 1118231568 M * IceTi LTNS???? 1118231586 M * IceTi Long time.... hmmm 1118231590 M * IceTi not# 1118231650 M * IceTi Long Time No See 1118231651 M * IceTi HA 1118231695 M * Bertl congrats! 1118231700 M * IceTi thanx 1118231802 M * IceTi öhm one question i have 1118231814 M * IceTi yoda style my language ;-) 1118231951 M * Bertl and probably answered it will be ;) 1118231992 Q * romke Ping timeout: 480 seconds 1118232032 Q * albeiro Ping timeout: 480 seconds 1118232135 J * albeiro ~albeiro@procyon.romke.net 1118232149 M * IceTi :-) 1118232154 M * Bertl IceTi: but not a great jedi yet -- ask you have to! 1118232170 J * romke ~romke@procyon.romke.net 1118232174 M * IceTi I think foget I have my question 1118232174 Q * romke Quit: 1118232179 M * IceTi a minute wait u must 1118232242 J * romke ~romke@procyon.romke.net 1118232400 J * rs ~rs@staff.lycos.fr 1118232475 J * romke_ ~romke@procyon.romke.net 1118232508 Q * romke_ Quit: 1118233277 Q * IceTi Quit: get satisfied! • :: ««« (Gamers.IRC) »»» www.gamersirc.net :: 1118233460 Q * monrad Quit: Leaving 1118234122 M * ruuth hi! I have mounted /usr/portage (ro,bind) and /usr/portage/distfiles (rw,bind) into my vserver. Now I want to install JOE - ist's already installed on the host. How can I configure my vserver that it tahes the precompiled version from the host instead of fetching the sources from the net? 1118234199 M * Bertl ruuth: sounds like a gentoo question to me? 1118234226 M * Bertl morning rs! 1118234239 M * rs morning bertl 1118234272 M * ruuth Bertl: :) ... yor're right - but I thought here's enought gentoo-knowledge it solve it - because in the gentoo-channels, they don't know vservers ;) 1118234322 M * Bertl ruuth: i.c. well, if some gentoo folks are around, and they know it, they'll probably answer 1118234454 M * ruuth Bertl: I hope so ;) 1118235535 M * ruuth Hullow: humm ... I did a fresh sync and wanted to emerge --update system ... and now there's a "Cannot find $EPATCH_SOURCE..." in baselayout-vserver-1.11.12-r4 1118235892 M * ruuth Hollow: humm ... I did a fresh sync and wanted to emerge --update system ... and now there's a "Cannot find $EPATCH_SOURCE..." in baselayout-vserver-1.11.12-r4 1118235938 Q * Bertl Ping timeout: 480 seconds 1118236097 Q * romke Quit: leaving 1118236140 J * romke ~romke@procyon.romke.net 1118237579 J * Bertl ~herbert@janus.mc.tuwien.ac.at 1118237760 M * romke Bertl: I still can figure out normal quota support in vserver-guest, have a minute to help? 1118237834 M * Bertl sure 1118237895 M * romke Bertl: What I need is using repquota and quota inside guest, no context tagging, standart system quota - is it possible? 1118237925 M * Bertl sure 1118237947 M * romke Bertl: with vroot? 1118237960 M * Bertl yep 1118237993 M * romke ls -alh /dev/vroot 1118238000 M * romke brw-r--r-- 1 root root 4, 0 Jun 8 15:17 0 1118238031 M * Bertl hmm, only one? 1118238058 M * romke Bertl: one per real partition - right? 1118238068 M * Bertl yep, that was on the host, no? 1118238081 M * romke yes, it is from host 1118238100 M * Bertl I just wonder why you have just a single device ... 1118238109 M * Bertl the default is 8 devices or so ... 1118238119 M * romke Bertl: manualy added by mknod 1118238132 M * Bertl hmm, why do you use devfs names then? 1118238154 M * Bertl well, anyway, it's probably fine ... 1118238179 M * romke Bertl: should it be /dev/vroot0 ? (hmm, it's only name difference) 1118238195 M * Bertl yeah, the canonical dev or udev name is vroot0 1118238231 M * romke Bertl: next, vrsetup /dev/vroot/0 /dev/hda3 1118238232 M * Bertl okay, now the setup ... 1118238287 M * romke Bertl: then cp -fa /dev/vroot/0 /vservers//dev/hdv1 1118238300 M * Bertl okay 1118238315 M * TheSeer Bertl: btw.. the nfs pr oblem vanished by switching to a different NIC 1118238330 M * romke Bertl: in guest, cat /etc/mtab gives: /dev/hdv1 / ufs rw,usrquota 0 0 1118238336 M * Bertl TheSeer: ah, so real hardware issues 1118238353 M * TheSeer yep.. 1118238353 M * Bertl romke: okay ... 1118238381 M * romke Bertl: in guest # quota 500 gives "quota: Quota file not found or has wrong format." 1118238390 M * Bertl what about the vrsetup? 1118238400 M * romke Bertl: after quotacheck new empty aquota.user file is created in / 1118238423 M * romke Bertl: vrsetup in guest? 1118238425 M * Bertl and where is your filesystem mounted? 1118238467 M * Bertl romke: no, vrsetup on the host, to configure the vroot device 1118238482 M * romke Bertl: 08>15:43 < romke> Bertl: next, vrsetup /dev/vroot/0 /dev/hda3 1118238501 M * Bertl ah, okay, missed that one 1118238512 M * Bertl where is hda3 mounted? 1118238518 M * romke Bertl: /dev/hda3 -> /vservers 1118238533 M * Bertl won't work, you need to mount it at /vservers/ 1118238545 M * Bertl otherwise the quota files will not be visible 1118238577 M * romke Bertl: so I can't have few vservers with that quota on one partition ? 1118238593 M * Bertl not without per context quota 1118238622 M * Bertl well, you can have the quota, but not inside the guests 1118238639 M * Bertl the quota tools insist on accessing the quota files 1118238642 M * romke Bertl: hmm, I'm thinking about hardlinking aquota.user - bad idea? 1118238653 M * Bertl probably bad idea 1118238666 M * Bertl it basically means that context A can modify the quota of context B 1118238796 M * romke Bertl: hmm, i don't care about that, I only need quota in guest to read usage, I set quotas on host 1118238813 M * Bertl well, you can 'try' then ... 1118238814 M * romke Bertl: guests are only for webhosting solutions 1118238835 M * romke Bertl: what about fs like xfs where no aquota files are created ? 1118238851 M * Bertl they have their own logic and tools ... 1118238866 M * Bertl you might try with that too .. it doesn't depend on vservers ... 1118238888 M * Bertl if they _can_ talk securely without or via the vroot device 1118238972 M * romke Bertl: currently I use awful method: on host in cron runs script parsing repquota and putting ussage to postgresql table, then all guests can read from there, but it sucks 1118239099 M * romke Bertl: anyway, thx for help 1118239262 M * Bertl hmm, why not make a service on the host which reads the quota? 1118239301 M * romke Bertl: same idea came to my mind, small daemon which will answer to alocation questions 1118239325 M * romke Bertl: but there's authentication issue for that 1118239374 M * Bertl yes? 1118239488 M * romke Bertl: hmm, I would have authenticate each connection and check for rights to check certain uids quotas 1118239541 M * Bertl I 'thought' root in the guest was trusted? 1118239578 M * romke Bertl: ok, is trusted, service based on unix socket it would be? 1118239618 M * Bertl for example 1118239620 M * romke Bertl: or on tcp socket? (does unix socket commucation work between contexts?) 1118239645 M * romke Bertl: or are there some limitations? 1118239647 M * Bertl tcp on local lowport is a good chouce ... 1118239653 M * Bertl *choice even 1118239780 M * romke Bertl: hmm, anyone from that machine (and all guests) can connect to it 1118239819 J * atsab ~as@lotes.vtu.lt 1118239939 M * romke Bertl: btw, guys in PLD prepared nice vserver-guest package replacements ( vserver-{dev,SysVinit,rc-scripts,basesystem}*.rpm ) :> 1118239964 M * Bertl nice 1118239971 M * Bertl welcome atsab! 1118240126 M * Hollow ruuth: please don't use baselayout-vserver yet.. 1118240133 M * Hollow it's for use with the new profile 1118240136 M * romke Bertl: I'll try with that service 1118240260 M * ruuth Hollow: Ok ... but why not? ... what's wrong? 1118240285 M * Hollow ruuth: the epatch error is an error in the ebuild, but the profile should not be used yet, it's to buggy 1118240325 M * ruuth Hollow: I see ... 1118240463 M * Hollow i'm currently working on it 1118240472 M * Hollow stage building etc 1118241473 M * ruuth Hollow: Ok ... I'll try again next week ... 1118241492 M * ruuth bye! 1118241571 M * Hollow i think the first stages should be ready this evening.. depends on teh compile time ;) 1118241609 M * Hollow first there will be basic i686 stages (1, 2 and 3) 1118241643 M * Hollow if you like to try the new profile tell me 1118242013 Q * ruuth Quit: Nettalk6 der Freeware IRC-Client 1118242310 Q * prae Quit: Execute Order 69 ! 1118242320 J * prae ~prae@ezoffice.mandriva.com 1118242337 Q * prae Quit: 1118242354 J * prae ~prae@ezoffice.mandriva.com 1118244021 J * cryo_ ~say@212.86.243.154 1118244069 Q * cryo Read error: Connection reset by peer 1118244069 Q * cryo_ Read error: Connection reset by peer 1118244080 J * cryo_ ~say@212.86.243.154 1118244621 M * Bertl okay, off for now ... back later 1118244628 N * Bertl Bertl_oO 1118244852 Q * prae Quit: Execute Order 69 ! 1118246501 Q * rs Quit: rs 1118247878 J * sim0n ~simon@vodsl-1560.vo.lu 1118247886 M * sim0n hey all 1118247918 Q * sim0n Quit: 1118248739 N * Doener` Doener|gone 1118248997 M * case yeah, tell me about configs. 1118249024 M * case 'flower page' makes my eyes cancer. 1118249754 J * erwan_ho ~erwan@konilope.dyndns.org 1118250192 M * TheSeer why don't we put the flower-page stuff into the wiki? 1118250205 M * TheSeer i mean.. without the cancer css ;> 1118251924 M * eyck what is flower-page? 1118251991 M * DaPhreak eyck: http://www.nongnu.org/util-vserver/doc/conf/configuration.html 1118252132 M * albeiro oh rotfl 1118253899 M * eyck I don't see no flowers 1118253912 M * eyck maybe dillo dosen't support css 1118254460 M * matti LOL 1118254470 M * matti This page rocks. 1118254471 M * matti ;] 1118254701 Q * TheSeer Ping timeout: 480 seconds 1118257411 Q * flock Ping timeout: 480 seconds 1118257505 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il 1118259047 Q * flock jupiter.oftc.net unununium.oftc.net 1118259047 Q * erwan_ho jupiter.oftc.net unununium.oftc.net 1118259047 Q * cryo_ jupiter.oftc.net unununium.oftc.net 1118259047 Q * atsab jupiter.oftc.net unununium.oftc.net 1118259047 Q * explasm__ jupiter.oftc.net unununium.oftc.net 1118259047 Q * id jupiter.oftc.net unununium.oftc.net 1118259047 Q * locksy jupiter.oftc.net unununium.oftc.net 1118259047 Q * mcp jupiter.oftc.net unununium.oftc.net 1118259047 Q * mugwump jupiter.oftc.net unununium.oftc.net 1118259047 Q * micah jupiter.oftc.net unununium.oftc.net 1118259047 Q * sladen jupiter.oftc.net unununium.oftc.net 1118259047 Q * SiD3WiNDR jupiter.oftc.net unununium.oftc.net 1118259048 Q * Snow-Man jupiter.oftc.net unununium.oftc.net 1118259048 Q * Beirdo jupiter.oftc.net unununium.oftc.net 1118259122 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il 1118259122 J * erwan_ho ~erwan@konilope.dyndns.org 1118259122 J * cryo_ ~say@212.86.243.154 1118259122 J * atsab ~as@lotes.vtu.lt 1118259122 J * explasm__ explasm@p549FF503.dip.t-dialin.net 1118259122 J * id ~id@relax-media.softwarezentrum.de 1118259122 J * locksy ~locksy@mrtg.sisgroup.com.au 1118259122 J * mcp ~hightower@wolk-project.de 1118259122 J * mugwump ~samv@210-54-92-184.ipnets.xtra.co.nz 1118259122 J * micah micah@micha.hampshire.edu 1118259122 J * Beirdo ~gjhurlbu@beirdo.usercloak.oftc.net 1118259122 J * SiD3WiNDR luser@bastard-operator.from-hell.be 1118259122 J * sladen paul@starsky.19inch.net 1118259122 J * Snow-Man ~sfrost@snowman.net 1118259122 T * xenon.oftc.net http://linux-vserver.org/ | latest stable 1.2.10, devel 1.9.5, 2.0-rc4, ng9.5 -- 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 ;) 1118261428 N * Bertl_oO Bertl 1118262561 N * Doener|gone Doener 1118262568 M * Doener evening folks! 1118262615 M * Bertl evening Doener! 1118262911 M * micah evening all 1118262911 M * Doener ah, i love it... finally all serious firefox bugs I experienced are gone... :) (well, one of them was sitting on my chair... ;) 1118262922 M * Doener hi micah 1118262966 M * micah I'm having a strange problem on a vserver machine host, the process nagios-statd-server cannot access disk information if it is run as nobody, but it can if run as root 1118262972 M * micah this is on the "host" not on a vserver 1118262986 M * micah and on other machines that are the same, except without vserver support, it can run as nobody 1118263001 M * Doener 'cannot access disk information' means? 1118263040 A * micah reads source 1118263067 M * micah hmm apparantly it does a df -P 1118263093 M * Doener 2.6 or 2.4? 1118263198 M * Bertl micah: permissions of proc maybe? 1118263211 M * Doener hm, if 2.4, are there any mounts below /vservers ? 1118263256 M * micah Doener: 2.6 1118263280 M * micah Bertl: this is what I am suspecting also, but it is on the host, not in any of the vservers 1118263288 M * micah Doener: there are some mounts below /vservers 1118263303 M * Doener you did not do "chmod 000 /vservers", did you? 1118263365 M * micah I did a while back 1118263370 M * mugwump morning all 1118263383 M * micah before I knew that you should do the chattr command instead, but perhaps I didn't un-do the chmod 000? 1118263401 M * Doener morning sam 1118263408 M * mugwump Bertl, I saw that unlink problem too though could never isolate it.. 1118263415 M * micah Doener: what should the permissions on /vservers be? 1118263424 M * Doener 755 should be ok... 1118263434 M * mugwump oh, 2.6 vs 2.4, n/m :) 1118263500 M * micah Doener: that solves part of the problem for sure! 1118263514 M * Doener good! 1118263514 M * micah I at least get it to report some information, although it is missing some now :P 1118263522 M * Doener not good! ;) 1118263532 M * Doener what is missing? 1118263557 M * micah Doener: it only prints info for /usr, but not for the other partitions 1118263582 M * micah I can work around that though 1118263672 M * Bertl mugwump: basically all permission issues have been fixed in latest versions 2.4 and 2.6 1118263697 M * mugwump ok, that *was* on 2.0-pre1 1118263745 M * mugwump no, I mean vs1.9.3-rc5 :) 1118264462 J * steve^ ~steve@user-2774.l6.c5.dsl.pol.co.uk 1118264469 M * steve^ hi all 1118264476 M * steve^ is the 2.6 tree stable enough for production use? 1118264478 M * Doener Bertl: hm, we're running out of xfs mount option flags if xfs folks go on like that ;) 1118264486 M * Doener steve^: lycos uses it in production 1118264557 M * Bertl steve^: hmm ... http://linux-vserver.org/VServer+Hosting 1118264616 J * yarihm ~yarihm@217-162-113-13.dclient.hispeed.ch 1118264953 M * Doener http://www.13thfloor.at/~doener/vserver/patches/patch-2.6.12-rc6-vs2.0-rc4.diff 1118264958 Q * steve^ Ping timeout: 480 seconds 1118265403 M * albeiro Bertl: today have seen flower page for the first time in graphical mode. it was... interesting experience ;p 1118265428 M * albeiro is anybody able to read it ? ;p 1118265448 M * Bertl yes, it has a lot of different style sheets ... 1118266458 J * Aiken ~james@tooax6-038.dialup.optusnet.com.au 1118266673 M * Doener Bertl: regarding the limits, what we do in case of a rss limit that is exceeded is to return VM_FAULT_OOM, which is not the oom killer, but just kill the process that generated the page fault, right? 1118266706 M * Bertl that's the idea ... 1118266708 M * Doener i.e. it does not select some task to be killed (including all threads using the same mm), but just kills the current thread 1118266741 M * Bertl hmm, you think we should kill the entire session? 1118266762 M * eyck hmm 1118266763 M * Doener not sure about that. 1118266770 M * eyck how should 'modern' configuration look like? 1118266786 M * Bertl like a dir tree ;) 1118266814 M * eyck i'm on 2.4 vs1.2.10.1, with sarge utils, and after newvserver I don't know what to do 1118266821 M * Doener thing is, my threaded memory hog eats memory, the threads hitting the limit get killed, but those not hitting the limit do not get killed, and the memory is still allocated 1118266833 M * Bertl eyck: newvserver is from debian tools 1118266846 M * eyck vserver gx enter ;capchroot: execvp("/bin/bash"): No such file or directory? 1118266856 M * eyck so what should I put in /etc/vservers ? 1118266872 M * Doener thus the vserver needs some kind of help from the host to kill the evil process, because everything started in the vserver gets killed 1118267056 M * Doener http://www.13thfloor.at/~doener/vserver/tools/oom-test-0.01.tar.bz2 1118267078 M * Doener start that one in a vserver that has only a hard rss limit, no vm limit 1118267226 M * Doener the "main" thread will not be killed and keep the memory allocated. you can't do much in the vserver then, because you hit the rss limit all the time 1118267313 M * eyck ~oh my gowd...... 1118267356 M * eyck I just saw flower page on mozilla.... 1118267409 M * albeiro oom-test: Mach-O executable ppc 1118267414 M * albeiro i love portable code ;p 1118267494 M * Bertl Doener: hmm ... well, that's the point, no? 1118267532 M * albeiro eyck: indeed it is unforgetable experience ;p 1118267567 M * Doener Bertl: hm? i had expected an oom killer like behaviour 1118267637 M * eyck can anyone show me how minimal 'modern' configuration should like like 1118267638 M * eyck ? 1118267641 M * eyck please? 1118267662 M * Doener eyck: vserver - build --help 1118267672 M * Doener then choose your options and do a skeleton build 1118267700 M * Doener i.e.: vserver testconf build -m skeleton 1118267732 M * Doener that will create a configuration using the given options and a minimal vserver directory (just /dev IIRC) 1118267931 M * eyck Doener: thanx. 1118267937 M * Doener you're welcome 1118267951 M * eyck hmm, newvserver is useless then? 1118267987 M * Doener I don't know what it does ;) but for my debian vservers I used the debootstrap build method from the tools 1118268007 Q * yarihm Quit: Leaving 1118268028 M * eyck newvserver calls debootstrap and creates /etc/vservers/$VS.conf, and that's it, i think. 1118268051 M * Doener then it seems to be useful for legacy tools only 1118268768 Q * erwan_ho Remote host closed the connection 1118268777 M * Bertl eyck: it seems newvserver is obsolete ... 1118268798 M * Bertl (well, at least that's what we agreed last time with ola) 1118268856 M * eyck ok. 1118268921 M * eyck why hostname doesn't change when I vserver test enter with new tools? 1118268936 M * Doener did you specify a hostname? 1118268943 J * Vinz Vinz@209.172.40.245 1118268951 M * Bertl welcome Vinz! 1118268954 M * Vinz hi there! 1118268983 M * Vinz last time i came here it was for a network issue ... i still have the same and i forgot the magic command :) 1118268993 M * eyck Doener: isn't 'test' supposed to be a hostname? 1118268998 M * Vinz i see all interfaces in my vservers ... i want to hide them 1118269011 M * Vinz it was something like hide_netif or something like that 1118269035 M * Vinz is there some way to automate that? i am using debian-vserver tools with legacy config 1118269074 M * Vinz I also remember that the capability id was 25 ... just don't remember how to change it 1118269089 M * Doener eyck: it's a vserver-name ;) the hostname is a separate option 1118269177 M * Vinz i tried to change that with vattribute ... but after that i got this error when trying to access the vserver : capchroot: chroot(): Operation not permitted 1118269287 M * Bertl Vinz: the first step into 'automating' that is throwing away the legacy tools 1118269302 M * Vinz hehe okay 1118269322 M * Bertl complain to your debian maintainer if he refuses to update the tools 1118269345 M * Bertl with new config/style it should be a) default to hide them, and b) configurable 1118269492 M * Bertl Vinz: and hide_netif is the right one ... 1118269505 M * Vinz great 1118269540 M * Vinz so first step would be to remove that capchroot: chroot(): Operation not permitted 1118269551 M * Vinz then i will migrate config from legacy 1118269593 M * Bertl Doener: the question is, do we really want to implement the 'every day changing' OOM behaviour for the guests? 1118269636 M * Bertl Doener: but maybe sending a kill to the taskgroup would be sufficient? 1118269667 M * Doener i'd say so... 1118269675 M * Doener what about an VM_FAULT_VX_OOM 1118269693 M * Doener and have that one do do_group_exit(SIGKILL) instead of do_exit(SIGKILL)? 1118269723 M * Doener hm, s/taskgroup/threadgroup/ ? (or is that even the same?) 1118269726 M * Bertl what about adjusting the do_exit()/VM_FAULT_OOM behaviour in general? 1118269739 A * Doener is confused by all those xyzgroups... 1118269766 M * Bertl yeah, tid, gid, pgid, pid ... 1118269786 M * Bertl session, group, taskgoup, threadgroup ... 1118269804 M * Bertl but the point is, if you say that is _unexpected_ behaviour 1118269825 M * Bertl then we should change it in general, not only for the vserver case, no? 1118269876 M * Doener i'm not yet sure what (semantically) causes handle_mm_fault() to return VM_FAULT_OOM in the non-vserver case, so... 1118270074 M * Vinz wow so much stuff to put in the new config format ... what were you thinking :) i will miss my one-file config 1118270112 M * Bertl the tools handle most of the config ... 1118270167 Q * Aiken Quit: Leaving 1118270302 M * Vinz what tools would i use to create the config for an already existing vserver 1118270356 M * Bertl the same, 'vserver' 1118270366 M * Bertl just move the 'guest' out of the way 1118270369 M * Medivh hi there ;) 1118270386 M * Bertl Vinz: then use the skeleton method to create a new one (with your config options) 1118270411 M * Vinz ok 1118270415 M * Bertl then move the guest back in place ... check the dev the newvserver utility always got that one wrong ... 1118270436 M * Medivh i was wondering if there's any electronics freaks around ;) 1118270444 M * Medivh quite off topic, but i dunno where else to ask... 1118270448 M * Bertl evening Medivh! hmm? 1118270464 M * Medivh thing is, i need to remove an ISA slot from a motherboard, i.e. unsolder it or something... wondering what is the best way to do it 1118270512 M * Medivh need the mobo working after, don't care about the isa slot :p 1118270556 M * Bertl for this you use something called 'entlötsaugpumpe' (no idea how to translate it ;) 1118270592 M * Bertl it's a suction/vacuum pump combined with a soldering iron (or in addition to one) 1118270619 M * Medivh yeah i know what you mean... 1118270632 M * Medivh no need to translate german for me after all *g* 1118270692 M * Medivh guess i should have one sitting around somewhere, good i didn't throw that defective same-model board away few weeks ago so i can try there first ;) 1118270756 M * Medivh just hope i can get it done without damaging the... err... leiterbahnen 1118270767 M * Bertl circuits ;) 1118270790 M * Medivh oh, yeah... stupid me ;) 1118270970 J * Aiken ~james@tooax6-038.dialup.optusnet.com.au 1118271199 M * Vinz is there a way to kill a vserver from the CTX^ 1118271200 M * Vinz ? 1118271228 M * Vinz the only thing i know from the running vservers is their CTX from vserver-stat ... I can't stop them by other means ... 1118271245 M * Vinz is there a kill --ctx type of command? 1118271302 M * Bertl vkill 1118271498 M * Vinz thanks Bertl! 1118271511 M * Bertl yw 1118271801 M * Vinz vserver name start won't start the network interface since i moved to the new config 1118271874 M * Vinz i started with vserver build -m skeleton and specified ip, hostname, etc. 1118271894 M * Bertl won't start means? 1118271895 M * Vinz but i took back the server image i had before when using the legacy tools 1118271928 M * Vinz means i don't see any new network interface with ifconfig and i don't see any other interface than eth0 from the vserver when doing ifconfig 1118271944 M * Bertl don't use ifconfig, it's outdated too 1118271953 M * Bertl use 'ip addr ls' 1118271980 M * Bertl if you want to have aliases (this is what ifconfig knows about) then you have to specify that in the config ... 1118272018 M * Vinz ok 1118272029 M * Vinz i don't really mind as long as the network interface works 1118272105 M * Aiken I'll admitt I was a confused as well. have not had a problem with networking. it just worked but still was expecting to see eth0:1 1118272114 M * Bertl it works quite fine ... if you want the alias, add an entry 'name' to the interface config (see FlowerPage for details) 1118272156 M * Bertl http://www.nongnu.org/util-vserver/doc/conf/configuration.html 1118272188 M * Vinz yeah, i know about that beautiful page :) 1118272192 M * Bertl (in this case, the name file would contain '1' ;) 1118272265 M * Bertl 'weedpage' is pretty nice actually ... except for the fixed header 1118272323 M * Vinz i always read it with styles disabled in firefox's web-developer toolbar :) 1118272400 M * Bertl de gustibus non est disputandum. 1118272434 M * Vinz creating a "name" file in /etc/vservers/id/interfaces/0 with content "1" doesn't change much ... still no alias 1118272452 M * Bertl did you restart the guest? 1118272494 M * Bertl what other files does the /etc/vservers/id/interfaces/0 contain? 1118272504 M * Vinz ip 1118272518 M * Vinz I restarted the guest many times 1118272525 M * Bertl that's all, just ip? 1118272527 M * Vinz I guess "su: Module is unknown" while it starts 1118272530 M * Vinz yes, just ip 1118272544 M * Bertl hmm, you didn't specify a device on creation? 1118272556 M * Bertl np, just add a file 'dev' 1118272556 M * Vinz yes but it's a level before 1118272567 M * Vinz in /etc/vservers/id/interfaces/dev 1118272595 M * Vinz okay i copied the file in ./0 1118272640 M * Bertl and btw, what tools do you use now? 1118272669 M * Bertl (what tool version I mean) 1118272686 M * Vinz vserver 0.30.204 1118272691 M * Bertl *sigh* 1118272700 M * Bertl please update to 0.30.207 1118272701 M * Vinz that's what comes with debian man :) 1118272723 M * Bertl well, debian is just released (sarg) and already ancient ;) 1118272730 M * Bertl *sarge even 1118272732 M * Vinz hehe 1118272739 M * Vinz might be in unstable then^ 1118272769 M * Bertl I still hope they update the tools and kernel soon 1118272793 M * Vinz okay the new tools are in unstable 1118272795 M * Vinz i will update to that 1118272799 M * Bertl excellent! 1118272841 M * Vinz restarted guess ... 1118272843 M * Vinz no alias 1118272850 M * Vinz still su Module error 1118272868 M * Bertl what is 'su Module error?' 1118272885 M * Vinz "su: Module is unknown" 1118272892 M * Vinz i get that when the image is booting 1118272904 M * Bertl aha? 1118272916 M * Vinz and i get that if i try "su" from inside the vserver 1118272943 M * Bertl aha? looks like the guest is damaged then? 1118272976 M * Vinz it was booting correctly from legacy 1118273011 M * Bertl I just checked: 1118273016 M * Bertl ls /etc/vservers/test3/interfaces/0/ 1118273017 M * Bertl dev ip prefix 1118273027 M * Bertl echo "1" >/etc/vservers/test3/interfaces/0/name 1118273038 M * Bertl vserver test3 start 1118273051 M * Bertl eth1:1 Link encap:Ethernet HWaddr 00:02:B3:23:D6:DA 1118273052 M * Bertl inet addr:192.168.3.1 1118273061 M * Vinz good, it's working for you :) 1118273074 M * Vinz should be working for me then 1118273080 M * Bertl vserver --version 1118273080 M * Bertl vserver 0.30.207 -- manages the state of vservers 1118273080 M * Bertl This program is part of util-vserver 0.30.207 1118273101 M * Bertl the entire guest was created with debootstrap 1118273112 M * Bertl (the debootstrap build option) 1118273148 M * Vinz ok 1118273152 M * Vinz i'll try that ... 1118273163 M * Vinz i get this error when stopping the vserver RTNETLINK answers: Cannot assign requested address 1118273178 M * Vinz i guess it's because the interface is not created so it can't remove it 1118273227 M * Bertl vserver vinz build -m debootstrap --hostname vinz --interface vinz=eth1:192.168.0.100/24 --context 42 -- -d sarge -m ftp://ftp.debian.org/debian 1118273235 M * Bertl (now building ;) 1118273321 M * Vinz hmm got it 1118273331 M * Vinz i forgot to specify /26 after my ipaddress 1118273343 M * Vinz it's now working 1118273348 M * Vinz although i can't su 1118273363 M * Bertl which is something which is broken in your guest 1118273382 M * Bertl I doubt you were able to do so before (in the legacy mode) 1118273413 M * Vinz i'll take a backup copy of the guest to make sure 1118273420 M * Bertl but in any case, check the permissions and flags of the vserver root dir, the debian tools always got that wrong 1118273480 M * Bertl ls -lda /vservers/test3 1118273481 M * Bertl drwxr-xr-x 20 root root 4096 May 24 06:29 /vservers/test3/ 1118273498 M * Bertl lsattr -d /vservers/test3/ 1118273499 M * Bertl ------------- /vservers/test3/ 1118273520 M * Bertl if you see 'i' or 000 permissions or so .. then it's wrong ;) 1118273567 M * Bertl and for the security: 1118273568 M * Bertl showattr -d /vservers/test3/. 1118273569 M * Bertl ---bui- /vservers/test3/. 1118273578 M * Bertl but 1118273579 M * Bertl showattr -d /vservers/test3/.. 1118273580 M * Bertl ---Bui- /vservers/test3/.. 1118273596 M * Bertl (not that important with namespaces, btw) 1118273608 M * Vinz i used a specific LVM logical volume for that guest 1118273623 M * Vinz it might be that the flags are not set 1118273716 M * Vinz /var/lib/vservers# lsattr -d /var/lib/vservers/ 1118273716 M * Vinz ----------------- /var/lib/vservers/ 1118273739 M * Vinz same for the guest 1118273809 M * Vinz I have ---bui- for the guest but ---Bui- for the /var/lib/vservers 1118273864 M * Vinz I did setattr --barrier on the guest's dir 1118274160 M * Bertl the guest dir should not have a barrier set 1118274178 M * Bertl the parent of the guest dir should ;) 1118274185 M * Bertl (see my example ;) 1118274240 M * Vinz ohh right, read to quick 1118274312 M * Vinz how do i unset the attribute then 1118274320 M * Bertl --~barrier 1118274380 M * Vinz copied from backup, restarted the vserver 1118274386 M * Vinz still have su Module error 1118274467 M * Vinz it has something to do with pam for sure ... 1118274471 M * Vinz but hard to say why 1118274491 M * Bertl strace? 1118274498 M * Bertl maybe check the logs? 1118274563 M * Vinz PAM unable to dlopen(/lib/security/pam_rootok.so) 1118274582 M * Bertl and, is that library there? 1118274589 M * Bertl is it accessible= 1118274590 M * Vinz nope 1118274623 M * Vinz but it's not in the backup i took on the working version 1118274631 M * Bertl but you're sure it worked before without it? 1118274687 M * Vinz it worked ... but i don't know how ... because this is very much needed 1118274722 M * Bertl well, I'd say, maybe it was 'loaded' but somebody removed it ... 1118274791 M * Vinz anyway, at least i know it's not vserver related 1118274885 M * Vinz thanks Bertl, you are always very helpful 1118274896 M * Bertl you're welcome!