1103500808 M * Bertl so basically all use ticks have been substracted but no tokens have been refilled 1103500823 M * Bertl okay, let's check the flags ... 1103500843 M * Bertl cat /proc/virtual/49179/status 1103500854 M * Thorsten UseCnt: 15 1103500854 M * Thorsten RefCnt: 13 1103500854 M * Thorsten Flags: 0000000202000010 1103500854 M * Thorsten BCaps: ffffffffd44c04ff 1103500854 M * Thorsten CCaps: 0000000000000101 1103500855 M * Thorsten Ticks: 0 1103500911 M * Bertl hmm, you missing flag 8 1103500914 M * Thorsten (I have to tell you that I do ps ax in the vserver because I can't find again what script I have to run) 1103500925 M * Thorsten ... that I can't do ... 1103500940 M * Thorsten /usr/lib/proc... or so 1103500948 M * Thorsten just in case that matters 1103500961 M * Bertl no, the script is called vprocunhide 1103500973 M * Bertl (no, it doesn't matter ;) 1103500978 M * Thorsten exactly I haven't run that, oh good 1103501004 M * Thorsten so flag 8 is missing. cat flag 1103501004 M * Thorsten sched_hard 1103501008 M * Thorsten is that right? 1103501013 M * Bertl vattribute --set --xid 49179 --flag ^8 1103501033 M * Thorsten cat /proc/virtual/49179/status 1103501033 M * Thorsten UseCnt: 15 1103501033 M * Thorsten RefCnt: 13 1103501033 M * Thorsten Flags: 0000000202000110 1103501033 M * Thorsten BCaps: ffffffffd44c04ff 1103501035 M * Thorsten CCaps: 0000000000000101 1103501036 M * Thorsten Ticks: 0 1103501049 M * Bertl okay, how does the sched look now? ;) 1103501050 M * Thorsten cat /proc/virtual/49179/sched 1103501051 M * Thorsten Token: 500 1103501052 M * Thorsten :-) 1103501056 M * Bertl much better ;) 1103501060 M * Thorsten :-))) 1103501075 M * Thorsten How does it come that the flag was missing? 1103501094 M * Bertl no idea, maybe it's not supported yet by the tools (some bug or so) 1103501106 M * Bertl maybe you are missing the magic 'newline' ;) 1103501131 M * Bertl try with: 1103501134 M * Bertl vattribute --set --xid 100 --flag sched_hard 1103501138 M * Bertl and vattribute --set --xid 100 --flag ~sched_hard 1103501148 M * Thorsten to read what? 1103501152 M * Bertl (where xid is your xid, btw you should use fixed xids) 1103501166 M * Thorsten to reach what? 1103501177 M * Bertl to set or delete that flag by hand 1103501195 M * Bertl Flags:0000000200000000 1103501197 M * Thorsten so you think my flag file is broken? 1103501200 M * Bertl # vattribute --set --xid 100 --flag sched_hard 1103501200 M * Bertl # cat /proc/virtual/100/status 1103501203 M * Bertl Flags:0000000200000100 1103501209 M * Bertl (so that works here) 1103501336 M * Thorsten Ok, I just tested it the vserver obviously can consume no longer all the cpu power! 1103501340 M * Thorsten Cool! 1103501352 M * Bertl it is ;) 1103501478 M * Thorsten I removed the linebreak in my flag file, it didn't changed anything (don't know if that was serious) 1103501512 M * Bertl well, not the removing, but a missing line break can cause some troubles with the tools 1103501529 M * Bertl you can check a vserver startup with --debug 1103501543 M * Bertl you should see the flag pop up in the final lines ... 1103501685 M * Thorsten Maybe it was all my fault ... wait 1103501756 M * Thorsten Is cat /proc/virtual/49185/status 1103501757 M * Thorsten UseCnt: 15 1103501757 M * Thorsten RefCnt: 13 1103501757 M * Thorsten Flags: 0000000202000110 1103501759 M * Thorsten right? 1103501795 M * Bertl looks good, just get a fixed context id ... 1103501892 M * Thorsten As it sill worked after I removed the linebreak I was suspicious, the file has to be named flags not flag. sorry! :-( 1103501905 M * Thorsten I'm such an i.... 1103501936 M * Bertl don't worry to much ... 1103501975 M * Thorsten I'm going to convert my gentoo test installation into a vserver in the next days, emerge world will be a good test for cpu hardlimits :-) 1103501979 M * Thorsten Thank Bertl 1103501983 M * Thorsten thanks 1103502030 M * albeiro i recomend you using binary packages wherever possible (generated from host system if also gentoo) 1103502036 M * albeiro hello btw :) 1103502115 M * Bertl hey albeiro! you _can_ now test ngnet ;) 1103502129 M * Thorsten Thanks albeiro but my host system is Debian, the gentoo installation is just for fun (and only a minimal one) 1103502224 M * albeiro Bertl: routing works ? great ! so i will not sleep tonight than ;p 1103502261 M * Bertl yeah, new version 'seems' to handle routing well ... 1103502278 M * albeiro we will see ;p 1103502293 M * albeiro i just can reverse patch previous version and patch this one ? 1103502315 M * Bertl well, yes and no, new patch is agains 1.9.3.11/2.6.10.rc3 1103502318 M * albeiro (not forgetting about delta first) 1103502328 M * albeiro ok, os i cna patch clean sources 1103502333 M * albeiro s/os/so/ 1103502346 M * albeiro s/cna/can/ 1103502348 M * albeiro ;p 1103502373 M * Bertl yeah?! ;) 1103502388 M * albeiro and advanced routing ? is it in plans ? 1103502435 M * Bertl plans yes, shouldn't be too hard, but for now 2 tables 1103502447 M * Bertl (i.e. local and main) 1103502807 M * Bertl uploaded the latest patch ... 1103502814 T * Bertl http://linux-vserver.org/ | latest stable 1.29, devel 1.3.9, 1.9.3, ng8.6 1103503098 M * albeiro now it just does not copmile with adv routing turned on 1103503135 M * Bertl yep, as I said --- for now just 2 tables (i.e. no advanced routing) 1103503143 M * albeiro yeah, i see 1103503170 M * albeiro (i can hardly see anything in this putty, i gues i should configure it better ;p_ 1103503670 Q * Thorsten Ping timeout: 480 seconds 1103504153 J * Thorsten ~Thorsten@dsl-084-057-012-013.arcor-ip.net 1103504328 M * albeiro back to linux, feeling much better ;p 1103504361 M * Bertl wb albeiro! ;) 1103504495 M * albeiro could you provede me links to necesary patches this time ? 1103504544 M * Bertl sure ... 1103504563 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.10-rc3-vs1.9.3.11.diff 1103504591 M * Bertl http://vserver.13thfloor.at/Experimental/NGNET/diff-2.6.10-rc3-vs1.9.3.11-ng8.6.diff 1103504605 M * albeiro k 1103504606 M * Bertl iptables stays the same 1103504624 M * Bertl (so if you already patched it everything is fine, if not): 1103504631 M * albeiro i did 1103504643 M * albeiro i have to read how it works internally (vserver iptables patch) 1103504648 M * Bertl okay, then I spare the link ;) 1103504656 M * albeiro sure ;p 1103504676 M * Bertl again keep the config as close as possible to the example config 1103504679 M * albeiro to make it suit my very big and nice firewall configuration 1103504713 M * albeiro i am trying to do keep it so 1103504734 Q * monrad Quit: Leaving 1103505626 M * complexho Bertl: All has gone well, I now have a shiny new LAMP environment... I really love the new utils :) 1103505637 M * complexho thanks for your help 1103505673 M * Bertl you're welcome! 1103505689 M * Bertl enjoy your new environment ;) 1103505700 M * complexho I will :) 1103505749 M * albeiro hm... 1103505777 M * complexho Nige the virtuozzo admin I told you about was round here tonight helping me set everything up, uncluding an 80 mile round trip to set up the ILO on the server so I could mess with the kernel more easily ;) 1103505792 M * complexho been quite an adventure :) 1103505817 M * Bertl ILO? 1103505824 M * albeiro aplied patch-2.6.10-rc3-vs1.9.3.11.diff and trying to do the same with ngnet but got error, cna i post 10 lines here ? 1103505831 M * complexho Lights-out management - remote console etc 1103505845 M * Bertl i.c. nice ... 1103505849 M * Bertl albeiro: sure ... 1103505860 M * albeiro flood comming ! 1103505862 M * albeiro violet linux-2.6.9 # patch -p1 --dry-run <../diff-2.6.10-rc3-vs1.9.3.11-ng8.6.diff 1103505865 M * albeiro patching file Makefile 1103505868 M * albeiro Hunk #1 FAILED at 1. 1103505871 M * albeiro 1 out of 1 hunk FAILED -- saving rejects to file Makefile.rej 1103505873 M * albeiro can't find file to patch at input line 16 1103505876 M * albeiro Perhaps you used the wrong -p or --strip option? 1103505878 M * albeiro The text leading up to this was: 1103505881 M * albeiro -------------------------- 1103505883 M * albeiro |diff -NurpP --minimal linux-2.6.10-rc3-vs1.9.3.11-ng8.5/drivers/net/vnet.c linux-2.6.10-rc3-vs1.9.3.11-ng8.6/drivers/net/vnet.c|--- linux-2.6.10-rc3-vs1.9.3.11-ng8.5/drivers/net/vnet.c 2004-12-15 15:23:25.000000000 +0100 1103505887 M * albeiro |+++ linux-2.6.10-rc3-vs1.9.3.11-ng8.6/drivers/net/vnet.c 2004-12-19 07:17:02.000000000 +0100 1103505890 M * albeiro :] 1103505922 M * Bertl hmm, indeed vnet is not part of the default patch, let me verify that ... sec 1103505956 M * Bertl oops, the patch is a delta to 8.5 1103505983 M * albeiro it would be logical 1103506021 M * Bertl okay, so please apply the ng8.5 patch 1103506030 A * albeiro does so 1103506050 M * Bertl then rename the patch to delta-ng8.5-ng8.6 1103506067 M * Bertl and apply it ontop of that ... sorry for the inconvenience 1103506100 M * albeiro no problem, just done it 1103506307 J * gotcha ~neqm@213.22.166.233 1103506320 M * Bertl hey gotcha! ;) 1103506335 M * gotcha hi there Bertl 1103506346 M * gotcha good night everyone 1103506366 M * albeiro good night gotcha ! 1103506539 M * gotcha I'am starting out with vservers and I must say the documentation isn't help much 1103506549 M * gotcha maybe I'am looking at the wrong documents 1103506559 M * Bertl probably ;) 1103506588 M * Bertl http://linux-vserver.org/Linux-VServer-Paper 1103506602 M * gotcha yes, I saw that one. good introduction and theory 1103506612 M * gotcha what about getting dirty? 1103506620 M * Bertl http://linux-vserver.org/alpha+util-vserver 1103506623 M * gotcha I've patched my kernel and installed the userland tools 1103506630 M * gotcha # vserver start 1103506630 M * gotcha Usage: /usr/local/sbin/vserver {start|stop|suexec|restart|condrestart|exec|enter|chkconfig|running|status} 1103506643 M * Bertl you need a vserver for that ... 1103506666 M * Bertl see the alpha util-vserver for creation examples 1103506699 M * gotcha oh alpha means 'alpha stage', I thought it was alpha arch 1103506702 M * gotcha dumb me 1103506726 M * Bertl well, the naming is confusing in many ways ... 1103506755 M * Thorsten And I hate it that the syntax is vserver NAME start and not vserver start NAME ;-) 1103506791 M * Bertl but I agree that the real 'starter' documentation is missing ... so if you feel like doing so, feel free to add a Newbie-Howto or something similar to the wiki 1103506861 M * gotcha oh always more work to me 1103506865 M * gotcha 8-) 1103506940 M * Bertl well, folks doing the coding have a lot of work too ... right? 1103506988 M * Thorsten good night @ all :-) 1103507017 M * Bertl night Thorsten! 1103507023 Q * Thorsten Quit: Leaving 1103507047 M * gotcha Bertl: extremely true 1103507146 M * rpetre Bertl: i pulled it off with alsa 1103507463 M * Bertl rpetre: so it works now? 1103507706 M * rpetre yes 1103507739 M * Bertl strange ... well, good for you ;) 1103507756 M * complexho Gotcha: I just went through the same pain tonight :) Got it all working now though... 1103507780 M * complexho a couple of gotchas along the way... 1103507795 M * Bertl lol 1103507808 M * complexho heh 1103507829 M * albeiro bug detected 1103507835 M * rpetre good night and thanks for all the help 1103507839 M * albeiro CC drivers/net/loopback.o 1103507839 M * albeiro CC drivers/net/vnet.o 1103507839 M * albeiro drivers/net/vnet.c: In function `vnet_open': 1103507839 M * albeiro drivers/net/vnet.c:162: error: structure has no member named `nfvnet' 1103507839 M * albeiro drivers/net/vnet.c: In function `vnet_close': 1103507841 M * albeiro drivers/net/vnet.c:181: error: structure has no member named `nfvnet' 1103507843 M * Bertl night rpetre! 1103507844 M * albeiro drivers/net/vnet.c: At top level: 1103507846 M * albeiro drivers/net/vnet.c:77: warning: 'vnet_vnet_xmit' defined but not used 1103507849 M * albeiro make[2]: *** [drivers/net/vnet.o] Error 1 1103507851 Q * rpetre Quit: nighty-night 1103507852 M * albeiro make[1]: *** [drivers/net] Error 2 1103507854 M * albeiro make: *** [drivers] Error 2 1103507857 M * albeiro ;p 1103507877 M * Bertl please upload your kernel config somewhere ... 1103507957 M * albeiro zeus.polsl.gliwice.pl/~albeiro/config 1103507986 M * albeiro it's quick'n'dirty, i'm not using development kernels like 2.6 are ;p 1103508116 M * Bertl hmm .. strange how did that sneak in? okay expect a diff in a few minutes 1103508155 M * albeiro k 1103508186 M * albeiro looks like something is missed 1103508206 M * Bertl no, those are changes from the next release, which do not work yet 1103508234 M * complexho night all thanks again Bertl :) 1103508240 M * Bertl night complexho! 1103508265 Q * complexho Quit: Leaving 1103508289 M * albeiro touch of The Future ;p 1103508384 M * Bertl http://vserver.13thfloor.at/Experimental/NGNET/delta-2.6.10-rc3-vs1.9.3.11-ng8.6-ng8.7.diff 1103508391 T * Bertl http://linux-vserver.org/ | latest stable 1.29, devel 1.3.9, 1.9.3, ng8.7 1103508435 M * albeiro thx 1103508443 M * albeiro that's what opensource means ! ;p 1103508546 M * albeiro 1st hunk failed for vnet 1103508557 M * albeiro i can correct it myself 1103508608 M * Bertl hmm .. how so? 1103508673 M * albeiro duno 1103508692 M * albeiro looks impossible 1103508726 M * Bertl such patch woes make me a little nervous ... 1103508761 M * Bertl because they usually originate from somehow different source trees ... 1103508857 M * gotcha # vserver teste start 1103508857 M * gotcha /usr/local/sbin/vserver-info: error while loading shared libraries: libvserver.so.0: cannot open shared object file: No such file or directory 1103508873 M * albeiro gr, xfs recompiles every time 1103508892 M * albeiro gotcha: add /usr/local/lib to /etc/ld.so.conf and issue ldconfig 1103508906 M * Bertl albeiro: (please check if you could revery, i.e. -R --dry-run) the full patch: http://vserver.13thfloor.at/Experimental/NGNET/diff-2.6.10-rc3-vs1.9.3.11-ng8.7.diff 1103508948 M * Bertl gotcha: yes, maybe doing ldconfig -v is enough ... 1103508993 M * albeiro Bertl: can not, gues what it thinks is not reversed - small part of vnet 1103509105 M * Bertl remove the drivers/net/vnet.c 1103509108 M * Bertl and do: 1103509123 M * gotcha albeiro, Bertl : yes, it worked. thank you. 1103509142 M * Bertl filterdiff -i '*/vnet.c' diff-2.6.10-rc3-vs1.9.3.11-ng8.7.diff | ( cd ; patch -p1 ) 1103509183 M * Bertl gotcha: you're welcome! 1103509186 M * albeiro i do not have command filterdiff :/ 1103509212 M * Bertl rpm -qf `which filterdiff` 1103509212 M * Bertl patchutils-0.2.24-1mdk 1103509242 M * albeiro emerge patchutils -va 1103509243 M * albeiro [ebuild N ] dev-util/patchutils-0.2.30 144 kB 1103509245 M * albeiro ;p 1103509258 M * gotcha it seems to be eating 99% of my cpu 1103509270 M * gotcha thats not good 1103509278 M * Bertl what is eating your cpu? ;) 1103509392 M * gotcha on top it shows 90.7% used cpu but the procs that are there are the ones before 1103509406 M * gotcha so I guess its the kernel 1103509421 M * albeiro Bertl: btw - i got amazed few minutes ago. full package of vim compilled with glibc is almost 1 MB in size, when the same package with uclibc is about 500 KB. nano (another small text editor) goes down from over 300 KB to 79 KB 1103509447 M * gotcha Bertl: after the vserver started 1103509473 M * Bertl well, meybe something inside the vserver does use up the cpu? check with vps 1103509634 M * albeiro Bertl: strange, this cmd by you gives me exactly nothing 1103509695 M * Bertl what do you mean by nothing? 1103509732 M * albeiro mayby another side - what would you expect ? 1103509798 M * Bertl it should recreate the previously deleted drivers/net/vnet.c 1103510237 M * gotcha I was wondering where can I read about after starting the vserver how does one interact with it via network 1103510258 M * albeiro Bertl: i thought so. it does not 1103510290 M * Bertl gotcha: if you configured a network address for the vserver, and you verified that it was started successfully (with vserver-stat), you can reach services over this network address 1103510302 M * Bertl you can also 'enter' the vserver and do stuff there 1103510329 M * Bertl albeiro: well, check the output of the filterdiff command ... 1103510363 M * Bertl (and the name of the patch, this is the _full_ patch) 1103510394 M * Bertl filterdiff -i '*/vnet.c' diff-2.6.10-rc3-vs1.9.3.11-ng8.7.diff | wc 1103510394 M * Bertl 618 1827 12842 1103510421 M * albeiro patching file drivers/net/vnet.c 1103510432 M * albeiro after a bit of thinking and tweaking ;p 1103510453 M * Bertl thinking is always appreciated ,) 1103510483 M * Bertl gotcha: does this answer your question? 1103510492 M * albeiro what else nice cmd are in this pkg ? i am going to look, filterdiff is nice 1103510522 M * albeiro many of them :) 1103510527 M * Bertl espdiff is very handy ... 1103510532 M * albeiro it seems that my life will be easier ;p 1103510543 M * albeiro Bertl: you mean ? some example ? ;p 1103510557 M * Bertl try: espdiff 1103510597 M * albeiro LOL 1103510613 M * albeiro albeiro@violet ~ % espdiff 1103510613 M * albeiro Please concentrate... 1103510613 M * albeiro No brainwave activity detected. Aborting. 1103510614 M * albeiro ups ;p 1103510631 M * gotcha Bertl: yes, thank you. but how does one enters the vserver? 1103510649 M * Bertl try vserver --help 1103510677 M * gotcha found it 1103510681 M * gotcha lame me 1103510715 M * gotcha # ifconfig eth0 up 192.168.0.2 1103510715 M * gotcha SIOCSIFFLAGS: Permission denied 1103510715 M * gotcha SIOCSIFADDR: Permission denied 1103510715 M * gotcha SIOCSIFFLAGS: Permission denied 1103510715 M * gotcha kafka:/# id 1103510717 M * gotcha uid=0(root) gid=0(root) groups=0(root) 1103510736 M * gotcha cool, this get a hacker a nice headache 1103510747 M * gotcha :-) 1103510793 M * Bertl with the stuff currently in testing, it will succeed ... but not work if you use the 'wrong' ips ;) 1103510843 M * gotcha why do say wrong ips? 1103510854 M * gotcha isn't this one allowed? 1103510905 M * Bertl yes, what I mean is, if you 'assign' 192.168.0.2 to a vserver, then the vserver root can only use that one, 192.168.0.3 would be 'wrong' 1103511095 M * gotcha but where, in the name of god, is this covered in the docs? 1103511128 M * Bertl what? the future development? nowhere ... 1103511177 M * gotcha I don't get it, that isn't implemented yet? 1103511275 M * Bertl the current implementation (the one you use) allows you to configure some ips on the host for each vserver 1103511288 M * Bertl and it forbids the vserver root to change any of them 1103511306 M * Bertl (that is documented in the paper and in many places) 1103511324 M * Bertl the currently developed ngnet will change that ... 1103511329 M * albeiro CC [M] net/ipv4/netfilter/ipt_addrtype.o 1103511329 M * albeiro net/ipv4/netfilter/ipt_addrtype.c: In function `match_type': 1103511329 M * albeiro net/ipv4/netfilter/ipt_addrtype.c:27: error: too few arguments to function `inet_addr_type' 1103511332 M * albeiro is it you ? 1103511333 M * gotcha ok, so I have only one interface, that's eth0, that I'am using already for the cable modem. also I made a eth0:0 alias at 192.168.0.1 1103511336 M * albeiro or just kernel fault ? 1103511346 M * gotcha and now I want 192.168.0.2 to be the vserver IP 1103511350 M * gotcha is that possible? 1103511374 M * Bertl albeiro: remove that ipt module 1103511377 M * albeiro k 1103511393 M * Bertl gotcha: yes, but it isn't required, you can use the 192.168.0.1 for the vserver too 1103511426 M * gotcha # ifconfig eth0 up 192.168.0.1 1103511426 M * gotcha SIOCSIFFLAGS: Permission denied 1103511431 M * gotcha how? 1103511449 M * Bertl just configure it when you create the vserver, or add the entries to the config 1103511467 M * Bertl (as I said, you do _not_ configure the interface from inside the vserver) 1103511488 J * DuckMaster ~Duck@dyn-83-152-128-166.ppp.tiscali.fr 1103511494 M * Bertl welcome DuckMaster! 1103511509 M * gotcha http://vserver.strahlungsfrei.de/tiki-index.php?page=ComplexSetting 1103511518 M * gotcha is this a hacked page or what? :-] 1103511560 M * Bertl well, it's a wiki ... 1103511595 M * Bertl http://vserver.strahlungsfrei.de/tiki-pagehistory.php?page=ComplexSetting&preview=12 1103511632 M * gotcha yeah 1103511637 M * gotcha will add the info again 1103511643 M * Bertl okay, thanks! 1103511673 M * albeiro CC [M] net/ipv6/af_inet6.o 1103511673 M * albeiro net/ipv6/af_inet6.c: In function `inet6_bind': 1103511673 M * albeiro net/ipv6/af_inet6.c:287: error: too few arguments to function `inet_addr_type' 1103511674 M * albeiro lol 1103511680 M * albeiro none ipv6 ? 1103511687 M * albeiro even in host system ? 1103511688 M * Bertl nope, not yet ... 1103511689 M * albeiro k 1103511731 M * Bertl but it is planned as soon as somebody with ipv6 knowledge helps me to test the basic stuff ... 1103511801 M * gotcha Bertl: I have some ipv6 experience if you need 1103511809 A * albeiro has such a knowledge 1103511818 M * albeiro and necesary ip space ;p 1103511823 M * albeiro /48 mask :) 1103511844 M * albeiro we will we will help you ;p 1103511846 M * albeiro (someday) 1103511847 M * Bertl gotcha: feel free to start testing the ngnet stuff ... 1103511865 M * Bertl (but I guess it's better to get used to linux-vserver stuff first) 1103511887 M * Bertl gotcha: but thanks for your offer ... 1103511891 M * gotcha Bertl: such an entrie to the config file would look like? I'am working on a newbie guide first 1103511892 M * gotcha :P 1103511920 Q * duckx Ping timeout: 480 seconds 1103511988 M * Bertl gotcha: on the alpha util-vserver page, there is a link to the 'flower page' 1103512002 M * Bertl (it basically describes all the various config files/options) 1103512023 M * Bertl but for a beginner it's easier to use the vserver build feature to get a feeling for that 1103512092 M * Bertl * vserver foo build -m apt-rpm --hostname vs.foo.org --netdev eth0 --interface 192.168.3.1/21 --context 42 -- -d fc1 1103512125 M * Bertl would assign the ip 192.168.3.1 in the subnet 192.168.0.0/21 1103512144 M * Bertl (and create a fc1 vserver for example) 1103512165 M * Bertl use the appropriate variations for debian and woody/sid 1103512186 M * Bertl you can also use the -m skeleton to avoid the vserver creation 1103512295 M * gotcha I already have one vserver, I don't want to build another one 1103512305 M * gotcha can't I just pass those args to vserver start? 1103512317 M * Bertl nope, but you can add them to the config 1103512384 M * Bertl /etc/vservers/vserver-name/interfaces 1103512402 M * albeiro btw 1103512404 M * albeiro WARNING: /lib/modules/2.6.10-rc3-vs1.9.3.11-ng8.7/kernel/net/ipv4/netfilter/ipt_VNET.ko needs unknown symbol vnet_get 1103512414 M * albeiro missing export_symbol in vnet.c 1103512439 M * Bertl hmm, maybe compile it into the kernel for now ... 1103512446 M * albeiro since you are not using modules in example config it does not show 1103512477 M * Bertl yeah, have to add it to the exports ... 1103512503 M * albeiro just did it 1103512694 M * gotcha Bertl: that flower css idea was really bad :P 1103512718 M * gotcha why not just don't use css, looks good and simple to me 1103512738 M * Bertl why not just select a different stylesheet /hint/ ;) 1103512984 M * gotcha select? there's no option for selection in the flower page 1103513040 M * Bertl it should be a 'feature' of your browser ... 1103513334 M * gotcha didn't know this feature 1103513342 M * gotcha we're always learning aren't we 1103513393 M * albeiro This is violet.inet (Linux i686 2.6.10-rc3-vs1.9.3.11-ng8.7) vc/4 2004-12-20 04:29:09 1103513397 M * albeiro Linux violet 2.6.10-rc3-vs1.9.3.11-ng8.7 #3 Mon Dec 20 04:15:22 CET 2004 i686 Pentium III (Coppermine) GenuineIntel GNU/Linux 1103513400 M * albeiro :) 1103513422 M * albeiro so... 1103513451 M * albeiro i have to recall where gentoo vserver howto was hiden ;p\ 1103513606 M * Bertl okay, so it is working so far? 1103513681 M * albeiro seems so 1103513685 M * albeiro now grepig logs 1103513991 M * mugwump Anyone know which kernel structure counts block device `usage' ? 1103514029 M * Bertl please define block device `usage' 1103514038 M * Bertl refcount? 1103514041 M * mugwump eg, in use by a filesystem 1103514091 M * mugwump Would you have to scan the namespace for vfsmounts and check their devnames perhaps? 1103514157 M * Bertl include/linux/device.h, struct device, kobject I'd say 1103514177 M * Bertl (but just an educated guess) 1103514204 M * mugwump mmm thanks that looks promising 1103514228 M * Bertl struct device * get_device(struct device * dev) 1103514228 M * Bertl { 1103514228 M * Bertl return dev ? to_dev(kobject_get(&dev->kobj)) : NULL; 1103514228 M * Bertl } 1103514232 M * Bertl yep 1103514266 M * Bertl (in drivers/base/core.c) 1103514272 M * albeiro Bertl: in guide i have found there is a little info about networking, i should not care about it i understand ? 1103514298 M * Bertl hmm, guide about what? 1103514324 M * Bertl you mean for testing with ngnet? 1103514432 M * albeiro yeah 1103514450 M * albeiro nevermind, going to sleep now 1103514463 M * albeiro will continue tommorow, much work to be done :) 1103514491 M * Bertl okay, have a good night then! 1103514502 M * Bertl and thanks for testing! 1103514615 M * gotcha Bertl: uff, this documentation really needs work for the newbie user 1103514667 M * albeiro have a good time ! 1103514725 M * Bertl gotcha: I'm the last one against extending/adapting the docu ... feel free to work on that ... 1103514767 M * Bertl (let me know when you have something, I'll gladly proof read it) 1103514788 M * gotcha to add an interface to a vserver I need to add the correct config entries right? 1103514808 M * Bertl basically yes 1103514860 M * gotcha can you, please, provide an example 1103514968 M * Bertl sec 1103515045 M * Bertl vserver XXXX build -m skeleton --force --hostname vs.foo.org --netdev eth0 --interface 192.168.0.2/24 --context 100 1103515080 M * Bertl assuming the vserver is named XXXX your ip should be 192.168.0.2/24 and the context is 100 1103515127 M * Bertl will write the following config: 1103515152 M * Bertl # ls /etc/vservers/XXXX/interfaces/ 1103515152 M * Bertl 0 dev 1103515173 M * Bertl # ls /etc/vservers/XXXX/interfaces/0 1103515173 M * Bertl ip prefix 1103515185 M * Bertl # cat /etc/vservers/XXXX/interfaces/dev 1103515185 M * Bertl eth0 1103515196 M * Bertl # cat /etc/vservers/XXXX/interfaces/0/ip 1103515196 M * Bertl 192.168.0.2 1103515205 M * Bertl # cat /etc/vservers/XXXX/interfaces/0/prefix 1103515205 M * Bertl 24 1103515219 M * Bertl (adding those by hand will do the same) 1103515269 M * gotcha # vserver teste start 1103515269 M * gotcha secure-mount: chdir("/tmp"): No such file or directory 1103515269 M * gotcha /usr/local/etc/vservers/teste/fstab:2:1: failed to mount fstab-entry 1103515269 M * gotcha find: var/run: No such file or directory 1103515269 M * gotcha fakerunlevel: open("/var/run/utmp"): No such file or directory 1103515274 M * gotcha after I do that I get 1103515275 M * gotcha this 1103515382 M * Bertl check the entries in /usr/local/etc/vservers/teste/fstab 1103515397 M * Bertl and compare them to your 'old' entries (it should have made a backup) 1103515634 M * gotcha why did you told me to build the vserver again when I already have one? 1103515660 M * Bertl I didn't tell you to build it again, did i? 1103515689 M * Bertl I just explained what the tools configure for an interface setup ... 1103515748 M * gotcha ok 1103515763 M * gotcha is they a way of rebuilding the config dir? 1103515788 M * gotcha i've lost it, but I don't want to retrieve all the debian stuff again, that I already have 1103515819 M * Bertl there should be a line like this: 1103515820 M * Bertl Renamed '/etc/vservers/.defaults/vdirbase/XXXX' to '/etc/vservers/.defaults/vdirbase/XXXX.~1103515808~' 1103515849 M * Bertl Renamed '/etc/vservers/XXXX' to '/etc/vservers/XXXX.~1103515808~' 1103515864 M * Bertl this is where your old config is stored ... 1103515871 M * gotcha no, there isn't, I removed it. the fstab has corrupted anyway 1103515882 M * gotcha is there other way? 1103515894 M * gotcha s/other/another 1103515912 M * Bertl how did you create the vserver in the first place? (cmd line please) 1103515925 M * gotcha vserver build -m debootstrap -- -d sarge 1103515964 M * Bertl then try ... 1103515992 M * Bertl vserver build -m skeleton --force --netdev eth0 --interface 192.168.0.2/24 --context 100 -- -d sarge 1103516034 M * Bertl (you might need to fixup the fstab by hand after that) 1103516051 M * gotcha ok, now I get the fstab thing 1103516056 M * gotcha how can I correct that? 1103516075 M * gotcha sorry for all the questions 1103516077 M * gotcha :] 1103516081 M * Bertl what does it contain atm ... 1103516104 M * gotcha none /proc proc defaults 0 0 1103516104 M * gotcha none /tmp tmpfs size=16m,mode=1777 0 0 1103516104 M * gotcha none /dev/pts devpts gid=5,mode=620 0 0 1103516154 M * Bertl hmm, looks good .. what do you get (error wise) 1103516174 M * gotcha # vserver teste start 1103516174 M * gotcha secure-mount: chdir("/tmp"): No such file or directory 1103516174 M * gotcha /usr/local/etc/vservers/teste/fstab:2:1: failed to mount fstab-entry 1103516174 M * gotcha find: var/run: No such file or directory 1103516174 M * gotcha fakerunlevel: open("/var/run/utmp"): No such file or directory 1103516278 M * Bertl ah, maybe the funny debian pathes are now broken ... 1103516326 M * Bertl let's verify that with vserver --debug teste start (please upload the output somewhere= 1103516329 M * Bertl ) 1103516371 M * gotcha nevermind, I'll download the stuff again 1103516381 M * gotcha no use in chasing this kind of stuff 1103516410 M * Bertl ok 1103516453 M * Bertl btw, if you have suggestions for improvements of the tools, please send them to the mailing list with cc to enrico ... 1103516494 M * gotcha ok 1103516665 M * gotcha i'am off now, thanks for all the help, you're a net-saint 1103516751 M * Bertl you're welcome! 1103517123 M * mugwump help early, help often. keeps 'em from coming back *really* confused 1103519535 Q * lilo Read error: Connection reset by peer 1103519748 J * lilo ~lilo@lilo.usercloak.oftc.net 1103520148 J * ntrs_ ntrs@Dardeene-68.188.50.87.charter-stl.com 1103520267 Q * pusling jupiter.oftc.net arion.oftc.net 1103520267 Q * ntrs jupiter.oftc.net arion.oftc.net 1103520267 Q * tchan jupiter.oftc.net arion.oftc.net 1103520421 J * tchan ~tchan@c-24-13-81-164.client.comcast.net 1103520447 J * pusling ~pusling@195.215.29.124 1103521841 J * ntrs ntrs@Dardeene-68.188.50.87.charter-stl.com 1103521923 M * Bertl night everyone ... cya tomorrow ... 1103521947 N * Bertl Bertl-zZ 1103522206 Q * ntrs_ Ping timeout: 480 seconds 1103524556 Q * sannes Read error: Connection reset by peer 1103527117 Q * dsanta_ Read error: Connection reset by peer 1103527145 J * dsanta ~santa@c68.190.156.105.roc.mn.charter.com 1103527590 Q * _are_ Quit: Disconnecting 1103530758 J * Thorsten ~Thorsten@dsl-084-057-079-015.arcor-ip.net 1103530961 M * meebey morning dudes 1103530965 M * meebey Bertl-zZ: ping 1103531078 M * Thorsten meebey, check when he went to bed ;-) http://vds.pas-mal.com/irclogs/vserver-log.20041220.txt 1103531200 M * meebey uuhh 1103531203 M * meebey little late 1103531218 M * meebey I have a strange problem and I am not sure if its vserver related 1103531521 J * sannes ~ace@home.skarby.no 1103531837 M * meebey anyhow I will give it a try 1103531849 M * meebey the problem is that some vservers with sarge report the harddisk is full 1103531854 M * meebey /usr/sbin/logcheck: line 140: cannot create temp file for here document: No space left on device 1103531857 M * meebey mail: /tmp/mail.RsXXXXGk6EtF: No space left on device 1103531860 M * meebey like here, in this case logcheck 1103531862 M * meebey Null message body; hope that's ok 1103531870 M * meebey I have the some on 2 other servers too using vserver 1103531890 M * meebey s/some/same/ 1103531914 M * meebey and there are too some programs, mostly crons reporting no space left on device 1103531932 M * meebey all servers have at least 2GB space, and inodes are not full 1103532361 M * meebey wuuut 1103532369 M * meebey seems like other programs get the same problem 1103532376 M * meebey System.IO.IOException: Win32 IO returned ERROR_HANDLE_DISK_FULL. Path: /usr/local/gas/logs/gas.log 1103532385 M * meebey wtf, ITS NOT FULL!111oneoneone 1103532387 M * Thorsten Maybe /tmp full? 1103532399 M * meebey Filesystem Size Used Avail Use% Mounted on 1103532400 M * meebey /dev/hdv1 110G 79G 27G 75% / 1103532415 M * meebey I can't reproduce it now, everything works perfect 1103532418 J * BWare ~bware@212.26.196.41 1103532424 M * meebey only on that cron batch it seems to happen 1103532426 M * meebey every day 1103532437 M * meebey maybe its the load? has linux a limit? 1103532574 Q * Thorsten Quit: Leaving 1103533369 Q * chand Quit: ircII EPIC4-2.0 -- Accept no limitations 1103536495 J * mboman ~michael@cm48.sigma230.maxonline.com.sg 1103539610 M * eyck no no, no no no no, no no no no, no no, there's no limits! 1103539703 M * Loki|muh hu? ;) 1103540283 Q * mboman Quit: One day I'll get that peer and reset HIS connection! 1103540453 J * mboman ~michael@cm48.sigma230.maxonline.com.sg 1103542403 J * _are_ ~are@gateway-dsl.lihas.de 1103542812 M * _are_ hi 1103542893 M * _are_ it occured 2 times now with 2.6.10-rc3-vs1.9.3.11 that /proc/net dssapeared from main server and all slaves -> no more entering the vservers, anyone else experienced this? and any other possibility to fix this except ... 1103542893 M * _are_ rebooting the main server? 1103543652 J * jsambrook ~jsambrook@aelfric.plus.com 1103544169 M * SiD3WiNDR eyck: :P 1103544355 J * Hollow ~bene@home.xnull.de 1103544380 Q * jsambrook Ping timeout: 480 seconds 1103545550 J * psychops psychops@lpc0220uns50.unimaas.nl 1103545591 M * psychops Hi, anyone here who can help on a bit of networking in a Vserver ? 1103545936 M * Loki|muh whats the problem? 1103546058 M * psychops Hi 1103546066 M * psychops well i am using devel version and tools 1103546084 M * psychops and want to add an extra ip to a vserver 1103546092 M * psychops so i added ./interfaces/0 1103546095 M * psychops and all the files 1103546102 M * psychops but that doesnt seem to work 1103546146 M * psychops i added interfaces/1 1103546153 M * Loki|muh what says "ip a s" inside this vserver? 1103546159 M * psychops ./interfaces/0 is allready there 1103546162 M * psychops hmm lemme check 1103546242 M * psychops sec i only have ip installed on root server 1103546250 M * psychops on root server it does showup btw 1103546254 M * psychops but not as an alias 1103546261 M * psychops inet 192.168.0.116/24 brd 192.168.0.255 scope global secondary eth0:vs6 1103546261 M * psychops inet 192.168.0.117/24 brd 192.168.0.255 scope global secondary eth0 1103546309 M * psychops what package supplies ip again ? 1103546321 M * Loki|muh iproute I guess 1103546362 M * psychops owh yea on gentoo ists iproute2 :D 1103546373 M * TheSeer "secondary eth0:vs6" looks pretty much 'alias' like to me.. 1103546401 M * psychops Yeah but second ip is 117 1103546412 M * psychops wich has no alias 1103546443 M * psychops inet 192.168.0.117/24 brd 192.168.0.255 scope global secondary eth0 1103546443 M * psychops inet 192.168.0.115/24 brd 192.168.0.255 scope global secondary eth0:vs5 1103546445 M * Loki|muh http://pastebin.com/131505 1103546449 M * psychops from within the vserver 1103546453 M * psychops so it does exist 1103546457 M * Loki|muh this is it for me outside the vserver 1103546531 M * Loki|muh and inside with only one ip configure inet 192.168.77.134/24 brd 192.168.77.255 scope global secondary eth0 1103546554 M * psychops hmm i think i am a dumbass 1103546568 M * psychops i forgot to init the device within the vserver 1103546734 M * psychops Yeah that fixed it 1103546828 M * TheSeer glad we could help ;-P 1103547206 M * psychops :) 1103547353 Q * ni Quit: leaving 1103547928 J * seving ~serving@213.186.170.110 1103547960 N * seving serving 1103548209 M * serving hi all 1103548225 M * serving my v stats shows 1103548227 M * serving 49154 4 10MB 1kB m00.00 m00.02 18m10.75 1103548248 M * serving no vserver name. how can I stop this vservr short of a reboot ? 1103548249 M * serving :) 1103548534 M * _are_ no expert, so might sound really stupid: if you vserver 'enter' your servers, the ctx-numer ist displayed 1103548562 M * _are_ so, stupid as I am I would enter each with a script, grep the ctx id and know who I am afterwards. :-> 1103548573 M * _are_ but experts have a real tool most likely. 1103548591 M * serving i dont understad. 1103548601 M * serving I have teh id 49154 1103548615 M * _are_ you have the names of your servers in general, right? 1103548633 M * serving yes. I can enter the vserver in question fine 1103548633 M * _are_ if I do vserver servername enter, I get the id displayed 1103548645 M * _are_ just hit 'vreboot' inside, I guess 1103548690 M * serving New security context is 49154 1103548696 M * serving which is wbat I started with 1103548719 M * _are_ now 'vreboot' should work 1103548730 M * serving i dont have vreboot 1103548762 M * _are_ you know the name of the vserver? 1103548772 M * serving yes. i do . 1103548792 M * _are_ what about vserer servername restart? 1103548798 M * _are_ what about vserver servername restart? 1103548853 M * serving that is the problem. it tells me the server is not running 1103548861 A * _are_ imagines the experts here falling under their tables as they can't hold back anymore at these approaches of mine. 1103548888 M * serving vserver-stats is not identifying the server by name and hence cant restart/stop it 1103548890 M * serving :) 1103548894 M * _are_ what happens if you start it, then? 1103548897 M * serving thanx _are_ 1103548904 M * serving it start starts fine. 1103548910 M * serving I have 4 instances so far 1103548911 M * serving LOL 1103548916 M * _are_ mine are not identified, either, I guess it is because i have the old conbfig file format 1103549045 M * serving i dont know about that. 1103549054 M * serving only this new server i am setting up is not identified 1103549060 Q * psychops Quit: 1103549074 M * serving the old one (from a year ago) is identified 1103549090 M * serving 49156 4 10MB 1kB m00.01 m00.02 28m57.65 1103549090 M * serving 49159 37 234MB 22kB m04.16 m02.07 m20.76 mx1 1103549380 M * serving ... 1103549381 M * serving .. 1103549382 M * serving . 1103549396 M * serving my vserver-stats shows 1103549401 M * serving 49154 4 10MB 1kB m00.00 m00.02 18m10.75 1103549412 M * serving no vserver name. how can I stop this vservr short of a reboot ? 1103549416 M * serving :) 1103549417 M * serving . 1103549444 M * Loki|muh with vtop you get the ctx-numbers 1103549456 M * Loki|muh try to identify the machine in this way ;) 1103549470 M * serving 49156 1103549478 M * serving now what ? :) 1103549546 M * serving vps ax shows 1103549554 M * serving 24506 49156 UNKNOWN ? 1103549643 M * Loki|muh vps ax | grep 49154 is empty? 1103550287 J * monrad ~monrad@213083190130.sonofon.dk 1103550421 Q * lilo Ping timeout: 480 seconds 1103550642 J * lilo ~lilo@lilo.usercloak.oftc.net 1103552174 M * gotcha (flood comming) 1103552178 M * gotcha # ./v_sshd 1103552178 M * gotcha exec /usr/local/sbin/chbind --ip eth0 /etc/init.d/sshd 1103552178 M * gotcha ipv4root is now 212.22.123.233 1103552178 M * gotcha chbind: execvp(): No such file or directory 1103552386 M * serving . 1103552422 M * Doener it should be '.../chbind --ip x.x.x.x....' IIRC 1103552446 M * Doener you can't bind to an interface 1103552480 Q * mboman Quit: One day I'll get that peer and reset HIS connection! 1103552939 J * mboman ~michael@cm48.sigma230.maxonline.com.sg 1103554299 N * Bertl-zZ Bertl 1103554454 M * Bertl morning folks! 1103554593 M * _are_ Hi Bertl 1103554630 M * _are_ it occured 2 times now with 2.6.10-rc3-vs1.9.3.11 that /proc/net dssapeared from main server and all slaves -> no more entering the vservers, anyone else experienced this? and any other possibility to fix this except ... 1103554630 M * _are_ rebooting the main server? 1103555035 Q * dsanta Remote host closed the connection 1103555039 J * dsanta ~santa@c68.190.156.105.roc.mn.charter.com 1103555399 M * Bertl _are_: sounds like you are removing the primary interface 1103555406 M * _are_ not really 1103555424 M * Bertl this might happen accidentially, if you do not assign an ip on an interface for the host (in the subnet the vservers are) 1103555441 M * Bertl (it's a deficiency of the ccurrent kernels) 1103555502 M * _are_ it is 10.1.0.0/16, all IPs 1103555516 M * Bertl okay, it's currently configured? 1103555528 M * _are_ yes 1103555601 M * Bertl ip addr ls | grep -v secondary | grep inet 1103555610 M * Bertl does that list any ip of a vserver? 1103555736 M * _are_ yes 1103555804 M * Bertl well, if you take down that vserver, all the others on the same interface are gone 1103555857 M * _are_ but eth0 != eth0:0 != eth0:1, right? 1103555865 M * Bertl nope 1103555880 M * _are_ and actually the directory /proc/net has not existed anymore 1103555899 M * _are_ nope as in they are same? 1103555909 M * Bertl yep eth0* is the interface 1103555930 M * Bertl eth0:0 is an alias on the same interface 1103555936 M * _are_ yes 1103555936 M * Bertl (so it will go down) 1103555956 M * _are_ eth0 never goes down, it is no vserver ip, only eth0:n are used for vservers 1103555969 Q * mboman Quit: One day I'll get that peer and reset HIS connection! 1103555981 M * Bertl the thing is, the network folks (linux netdev) say: there can only be _one_ primary address, all other addresses are secondaries 1103556003 M * Bertl and they further say, if the one primary address goes down, all the secondaries have to go too .. 1103556019 M * Bertl ip addr ls 1103556030 M * Bertl will list the interfaces (numbered) 1103556047 M * Bertl and mark secondaries as 'secondary' 1103556101 M * Bertl the workaround is simple in your case ... just take an unused address in 10.1.x.x and assign it to eth0 at bootup 1103556128 M * _are_ hmm, messed round a lot in network it looks like. will reboot, then test again 1103556466 M * Doener Bertl: hm, but that issue won't make /proc/net disappear, does it? 1103556480 M * Bertl well, usually not ... 1103556481 M * Doener s/does/will/ 1103556599 M * _are_ hmm, now is set up the virtual interfaces for all vservers in /etc/network/interfaces at boot time and whoops, /proc/net gone from the start 1103556632 M * Bertl on the host, ls /proc lists no 'net' ? 1103556642 M * _are_ it does 1103556660 M * Bertl but? 1103556665 M * _are_ no, it doesn't. 1103556683 M * Bertl what does it list on the host 1103556683 M * _are_ ls -la /proc | grep net 1103556684 M * _are_ ls: /proc/net: No such file or directory 1103556684 M * _are_ dr-xr-xr-x 2 root root 0 Dec 20 16:26 virtnet 1103556700 M * _are_ seems the fs entry is there, but no access or the like 1103556714 M * Doener hm, that sounds like proc-security IIRC... 1103556731 M * Bertl no, proc security would hide them ... 1103556733 M * Doener (that part) ls: /proc/net: No such file or directory 1103556746 M * Bertl _are_: try the following: 1103556766 M * Bertl chcontext --ctx 1 ls -la /proc/ | grep net 1103556826 M * Doener i remember having seen this before, dunno if it was proc or devpts... 1103556852 M * _are_ # chcontext --ctx 1 ls -la /proc/ | grep net 1103556852 M * _are_ dr-xr-xr-x 5 root root 0 Dec 20 16:26 net 1103556856 M * Bertl I have a few ideas, but let's try to check some things first 1103556871 M * _are_ i used vprocunhide, if that helps 1103556877 M * Doener doener ~ # chcontext --ctx 100 ls /proc/net/ 1103556877 M * Doener New security context is 100 1103556877 M * Doener ls: /proc/net/: No such file or directory 1103556884 M * Bertl okay, now let's check with lsxid 1103556885 M * Doener with hidden /proc/net ;) 1103556907 M * Doener argh, i'm stupid... 1103556928 M * Bertl _are_: where is your patch from, btw? 1103556935 M * Bertl (the kernel patch) 1103556959 M * _are_ 2.6.10-rc3-vs1.9.3.11 1103556967 M * _are_ the one you suggested me to try few days ago 1103557022 A * Doener will get some coffee before trying the next stupid thing... ;) 1103557124 M * _are_ guess i got it from http://vserver.13thfloor.at/Experimental/ 1103557131 M * Bertl sec 1103557290 M * albeiro :] 1103557336 M * Bertl _are_: okay, what does uname -a on that host return? 1103557351 M * _are_ Linux master1a 2.6.10-rc3-vs1.9.3.11 #1 SMP Thu Dec 16 20:48:31 UTC 2004 x86_64 GNU/Linux 1103557369 M * Bertl okay, good, could you try: 1103557369 Q * sannes Read error: Connection reset by peer 1103557377 M * Bertl lsxid /proc/net 1103557396 M * _are_ # lsxid /proc/net 1103557396 M * _are_ lstat(): No such file or directory 1103557427 M * Bertl what about 1103557429 M * Bertl chcontext --ctx 1 lsxid /proc/net 1103557445 M * _are_ # chcontext --ctx 1 lsxid -d /proc/net 1103557445 M * _are_ New security context is 1 1103557445 M * _are_ vc_get_iattr(): Function not implemented 1103557445 M * _are_ !!ERR!! /proc/net 1103557463 M * Bertl hmm .. okay ... 1103557467 M * _are_ without -d i get the content, aka many 'Function not implemented' lines 1103557477 M * Bertl yeah, forgot the -d 1103557523 M * Bertl vatrrsetattr --~hide /proc/net 1103557525 M * Bertl arg 1103557529 M * Bertl setattr --~hide /proc/net 1103557541 M * Bertl then 1103557550 M * Bertl lsxid -d /proc/net 1103557553 M * _are_ # setattr --~hide /proc/net 1103557553 M * _are_ lstat(): No such file or directory 1103557571 M * _are_ # chcontext --ctx 1 setattr --~hide /proc/net 1103557571 M * _are_ New security context is 1 1103557612 M * Bertl hmm, now I wonder why setattr tries to lstat the path? 1103557631 M * _are_ btw: bash_completion finds /proc/net, ls doesn't 1103557654 M * Bertl okay, does the lsxid now work on /proc/net ? 1103557678 M * _are_ no, neither normall, nor in context 1 1103557771 M * Bertl okay, might be an x86_64 specific issues ... let's get the old tools ... 1103557780 M * Bertl sec 1103558067 Q * Hollow Quit: Leaving 1103558099 M * Bertl the network today is in excellent shape ... can not even realiably reach google :/ 1103558382 M * serving . 1103558403 M * serving hi Bertl 1103558443 M * Bertl hey serving! 1103558444 M * serving am having a problem stoping a vserer. It starts fine but vserver-stats shows no name for it. 1103558468 M * serving how can I kill it sing its id ? 1103558484 M * serving 49154 4 10MB 1kB m00.00 m00.02 18m10.75 1103558499 M * serving sing/using 1103558502 M * Bertl sounds like a legacy server with new tools (and one with a dynamic xid too ;) 1103558518 M * serving yes. it is 1103558538 M * Bertl well, using dynamic xids is a bad idea ... (I always say that) 1103558556 M * serving about 11 month old technology 1103558558 M * serving :) 1103558564 M * Bertl you now can 'guess' which server that was, and try to stop it 1103558590 M * serving well . it is a test box. 1103558601 M * serving I know which server I started 1103558610 M * Bertl then stop that one ;) 1103558613 M * serving there are 4 instances of it btw 1103558628 M * serving it says it is not running 1103558651 M * Bertl then something is confused ... somehow 1103558661 M * serving right 1103558685 M * serving is there a way to kill it using the id ? 1103558732 M * Bertl chcontext --xid kill ... 1103558735 M * Bertl or vkill 1103558805 M * Bertl _are_: try to find the vproc tools 1103558820 M * Bertl (I have almost no internet conenctivity atm) 1103558958 M * serving vkill 49156 1103558958 M * serving Process id 49156 not found. 1103558978 M * serving that id 49156 was taked from vserver-stats 1103558989 M * Bertl _are_: http://vserver.13thfloor.at/Experimental/TOOLS/vproc-0.01.tar.bz2 1103558991 M * _are_ Berl: vproc tools is an extra package? 1103559000 M * _are_ ok, it is ;) 1103559002 M * Doener serving: you need the process id, not the context is 1103559006 M * Doener s/is/id/ 1103559008 M * Bertl _are_: you have to adapt the syscall 1103559015 M * serving Oh dumb me 1103559028 M * Bertl nope, sec 1103559108 M * Doener or use: vkill --xid 1103559127 M * _are_ Bertl: got and compiled it, but what to do with it? and which syscall? 1103559129 M * Doener (with alpha tools) 1103559253 M * Bertl serving: yep, that's the way to do it ;) 1103559311 M * Bertl hmm, nope, the vproc is using the ioctl 1103559378 M * Bertl _are_: okay, we'll have to write a new hack tool for that ... will take a few minutes 1103559427 A * _are_ whistles innocently. 1103559455 M * _are_ sorry, I don't really understand the problem, yet, so thanks for any tools ;) 1103559484 M * serving tahnx guys . I killed them all 1103559531 M * serving can I put a vserver in a dir other than /vservers ? is it documented somewhere ? 1103559717 M * Doener serving: alpha tools? 1103559762 M * serving no. still using toold form about 10 month ago 1103559769 M * serving toold/tools 1103559862 M * serving 0.29 to be exact 1103559881 M * Doener IIRC they need to have the same path for all vservers 1103559897 M * serving yup. that is the way it was then too 1103559911 M * serving thanx Doener :) 1103559942 M * Doener alpha tools allow to put them in different locations, since they keep symlinks to the vservers' roots instead of having the path hardcoded somewhere 1103559949 M * serving can I mount dirs under /vservers ? would a ln -s do ? 1103560048 M * Doener symlinks are a bad idea i'd say, but (bind-)mounts should be fine 1103560373 J * Hollow ~bene@home.xnull.de 1103561134 M * Bertl welcome Hollow! 1103561616 M * Hollow hey Bertl 1103561655 M * Hollow i just tried the schedule howto from the mailing list.. i still get the segfault with vsched :/ 1103561669 M * _are_ Bertl: any progress on that tool you do, that might help me and that I have no idea about what it might do? ;) 1103561682 M * Bertl yes, still working on it ... 1103561709 M * Bertl _are_: talk to Hollow, maybe you can help him with vsched 1103561723 M * _are_ uhm, not tried vsched as of now 1103561738 M * _are_ but compiled it in, so can try as well 1103562219 M * _are_ ok, set it up, but seems not to work like expected, however, no segfalts here 1103562234 M * Bertl both archs are x86_64, right? 1103562262 M * Hollow no, x86 here 1103562300 M * Bertl hmm ... thought that was x86_64 ... okay, my fault 1103562326 M * Bertl Hollow: did you try the cvs version? 1103562341 M * _are_ Hollow: indeed, get a segfaulot, too 1103562362 M * Hollow Bertl: i looked at cvs, but there is no change for about 2 months in vsched.c 1103562378 M * Bertl okay, and where does it segfault? 1103562392 M * Hollow sec... 1103562392 M * Bertl (do we know that yet?) 1103562394 M * _are_ ok, i get the segfault, *but* the values are set anyway 1103562413 M * Bertl yes, I remember that this was posted on the ml ... 1103562420 M * _are_ Bertl: strace any useful? 1103562433 M * Bertl what command do you use to get the segfault? 1103562451 M * _are_ vsched --xid 49161 --fill-rate 20 --interval 100 --tokens 500 --tokens-min 0 --tokens-max 1000 --prio-bias 0 1103562501 M * Hollow yep.. they are set here too 1103562509 M * Hollow stoned root # vsched --xid 111 --fill-rate 20 --interval 100 --tokens 600 --tokens-min 0 --tokens-max 500 --prio-bias 0 1103562509 M * Hollow Segmentation fault 1103562509 M * Hollow stoned root # cat /proc/virtual/111/sched 1103562509 M * Hollow Token: 500 1103562509 M * Hollow FillRate: 20 1103562511 M * Hollow Interval: 100 1103562513 M * Hollow TokensMin: 0 1103562515 M * Hollow TokensMax: 500 1103562517 M * Hollow PrioBias: 0 1103562519 M * Hollow cpu 0: 29942 4618 14451 1103562523 M * _are_ i altered all values, they don't influence the segfault 1103562585 M * Bertl add -- true 1103562601 M * Bertl (as last part) 1103562619 M * _are_ works 1103562659 M * Hollow yep 1103562680 M * Bertl fine, please make a bug report on the ml, cc to enrico 1103562705 M * _are_ Hollow: you do it or I do it? 1103562978 M * _are_ almost done... 1103562990 M * Bertl me too ;) 1103563016 M * _are_ sent. 1103563023 M * _are_ so any more reports are double ;) 1103563151 M * _are_ for vserver interface name: the virtual interfaces are generated automatically when the vservers are started, is it good to specify them in the mainserver for bootup, too, or is this contraproductive? 1103563186 M * Bertl please rephrase ... 1103563226 M * _are_ well, while no vserbers are started, is it good to have all interface aliases with the vserver ips specfied? if they are not specified, they are generated by vserver start 1103563284 M * Bertl depends on your config 1103563305 M * Bertl if your config says, create alias or add ip, then it would be bad to have them set before 1103563329 M * Bertl OTOH, if your config says, don't create just use, then _not_ having them would be bad 1103563351 M * _are_ ahhhhhhhhh 1103563365 M * _are_ guess i go through that and verify it is consitent 1103563464 M * Bertl okay, tool is done, I'll test it now 1103563581 M * _are_ hmm, got lost in manuals (again): where is the old config format described? (and while we are at it: where the new one?) 1103563627 M * Bertl flower page has the new one (or the xml file inside the tools' source) 1103563948 M * Bertl _are_: http://vserver.13thfloor.at/Experimental/TOOLS/vxid-0.02.tar.bz2 1103563973 M * Bertl you have to adjust the syscall id in vserver.h to match x86_64 syscall number 1103564001 M * _are_ cc -g -Wall -O2 -I. -o vxid vxid.c 1103564001 M * _are_ In file included from vxid.c:22: 1103564001 M * _are_ vserver.h:6:1: warning: "__NR_vserver" redefined 1103564016 M * Bertl just comment it out ... for a test 1103564034 M * Bertl vxid -V 1103564046 M * Bertl should report the interface version if everything is working fine ... 1103564055 M * _are_ oh, and syscall number: what is this and how do I find the correct one? 1103564057 M * albeiro http://home.xnull.de/work/gentoo down as always... 1103564077 M * Bertl it looks to me like your setup already has the syscall number 1103564092 M * Bertl (don't know why, but probably because your glibc headers are broken ;) 1103564106 M * _are_ # ./vxid -V 1103564106 M * _are_ version: 0001:0024 1103564111 M * Bertl excellent! 1103564122 M * albeiro anybody got that site mirrored ? 1103564135 M * Bertl _are_: now make /proc/net 'disappear' and do: 1103564144 M * Bertl vxid /proc/net 1103564181 M * albeiro google has ;p 1103564210 M * _are_ # ./vxid /proc/net 1103564210 M * _are_ set_iattr:: No such file or directory 1103564210 M * _are_ xid=-1073742680, flags=0x0000007f, mask=0x00000002, /proc/net 1103564267 M * Bertl okay, that is probably a copy paste error ... 1103564309 M * Bertl in line 174 (vxid.c) it should read perror("get_iattr:") 1103564310 M * _are_ of mine? of kernel developer? of vserver patch developer? of vserver-util deveoper? 1103564339 M * Bertl it's my copy & paste error!! exclusively mine!!! 1103564360 M * daniel_hozac why the :? doesn't perror output : strerror(errno)? 1103564360 M * _are_ perror("set_iattr:"); 1103564373 M * Bertl but it doesn't change the fact that the get fails ... 1103564374 M * _are_ there is a ; appended (afaik necessary) 1103564392 M * Bertl yes, just replace the "s by "g 1103564427 N * ravenedge click 1103564446 M * _are_ # ./vxid /proc/net 1103564446 M * _are_ get_iattr:: No such file or directory 1103564446 M * _are_ xid=-1073742680, flags=0x0000007f, mask=0x00000002, /proc/net 1103564562 J * sannes ~ace@home.skarby.no 1103564644 M * Bertl welcome sannes! 1103564656 M * albeiro Bertl: to build skeleton cmd like this is sufficient ? vserver gentoo-template build -m skeleton --hostname gentoo-template --flags lock,nproc --initstyle gentoo 1103564702 M * Bertl if initstyle gentoo is define, and you _really_ want those flags then yes 1103564727 M * albeiro i gues i may not want them ? 1103564729 M * _are_ argh, missing a date, back in 4h 1103564737 N * _are_ are|afk 1103564754 M * albeiro lock,nproc -> what does they mean ? ;p 1103564783 M * albeiro of course i am using ngnet, so i do not need to add any networking options to it ? 1103564829 M * Bertl lock is a legacy flag which disables context migration, and nproc is a legacy flag too which doesn't mean anything right now 1103564979 M * albeiro so i od not want them :) 1103564994 M * Bertl usually not 1103565068 M * albeiro http://www.google.pl/search?q=cache:EHegjGFKUfsJ:home.xnull.de/work/gentoo/vserver/guide/++vserver++xnull++gentoo&hl=pl&ie=UTF-8&inlang=en 1103565068 M * albeiro google copy of gentoo guide 1103565068 M * albeiro that's what i am taking ideas from 1103565256 M * Bertl hmm, why not from http://home.xnull.de/work/gentoo/vserver/guide/ 1103565299 M * Bertl anyway, no idea why they have those flags set 1103565314 M * albeiro that site does not work for me 1103565345 M * Bertl best to send an email to # bb[at]xnull[dot]de 1103565354 M * Bertl and ask why they use those flags ... 1103566051 M * Bertl okay, back later ... 1103566056 N * Bertl Bertl_oO 1103566618 J * Thorsten ~Thorsten@dsl-084-057-027-201.arcor-ip.net 1103567430 J * mikmu hadge@h64-5-199-35.gtcust.grouptelecom.net 1103567484 M * mikmu Hey there, no one here working with the -mm kernel patches, are they? I know that the 1.9.3 doesn't patch to the -mm kernels, but I was looking into using reiser4, which is included in -mm 1103567551 M * mikmu oh, and what's ng8.7? Haven't been around in a bit 1103569756 M * matti Eh. 1103569760 P * matti 8-X 1103570558 M * serving . 1103570599 M * serving I am starting a new vserver but its name is visible in vserver-stat and it can not be shutdown. 1103570614 M * serving which fils are needed for the name to be working properly ? 1103570647 M * serving I have vs.sh vs.conf . I added hosts fine 1103570772 M * serving is NOT visible 1103570776 M * serving :)) 1103571243 Q * Hollow Quit: Leaving 1103573909 Q * mikmu Quit: 1103575186 Q * serving Read error: Connection reset by peer 1103576817 M * albeiro strange 1103576828 M * albeiro i have no idea if i have vserver started ;p 1103576831 M * albeiro funny 1103577591 Q * sannes Read error: Connection reset by peer 1103578053 Q * are|afk Quit: Disconnecting 1103579702 N * Bertl_oO Bertl_zZ 1103580500 J * are|afk ~are@dsl-213-023-050-156.arcor-ip.net 1103581497 M * are|afk hi 1103581504 N * are|afk _are_ 1103582148 J * serving ~serving@213.186.170.110 1103583615 J * ccooke ~ccooke@spc1-walt1-4-0-cust221.asfd.broadband.ntl.com 1103584564 J * sannes ~ace@home.skarby.no 1103585616 N * _are_ are|afk