1132446311 M * Bertl ag-: you reported the df issues with loop devices, right? 1132446345 M * daniel_hozac yes. 1132446415 M * Bertl ag-, daniel_hozac: how do you test? 1132446421 M * ag- Bertl: yes, i'm gonna test them further 1132446906 M * ag- I have a plain ext2-formatted file. When I issue losetup file /dev/loop0 and mount -otagxid /dev/loop0 /var/lib/vservers/test/, then everything is fine. When I issue mount -oloop,tagxid file /var/lib/vservers/test/, then df reports 0 on available space. 1132446957 M * ag- In the first case, with /dev/loop0 directly mounted, vdlimit done, df also reports 0 on available space. 1132447008 M * ag- I did proper chxid -Rc test on the filesystem too in both cases. 1132447038 M * ag- It's something weird I'm gonna test on another non-debian sarge setup... 1132447129 M * ag- i'm afraid of starting to hit the bug "sarge is growing old" :) 1132447783 M * ag- actually, it's ext3-formatted 1132448270 M * Bertl ag-: I'd say your mount command gets the mtab wrong, and df relies too much on that ... 1132448302 M * Bertl ag-: please make a copy of the mtab in both cases and upload them 1132448804 M * sizo n8 1132449177 M * Bertl night sizo! 1132449997 Q * logger Remote host closed the connection 1132450483 Q * yarihm Quit: Leaving 1132450572 M * Hollow ok, off to bed, cu tomorrow 1132450803 M * Bertl night Hollow! 1132451698 J * Cru ~mindwarp@bastardrouterfromhell.e.de.wahlich.com 1132451702 M * Cru re 1132451777 M * Bertl welcome Cru! 1132451919 M * Cru Bertl: be honest, do you have some kind of "Press F1 to say welcome to " script? ;) 1132451975 M * Cru just for interest ;) 1132452567 M * Bertl nope, sorry ... nothing automated yet 1132452589 M * Bertl but I'm using tab completion most of the time ... 1132452593 M * Cru just looked like anyway ;) 1132452630 M * Bertl yeah, well, in the beginning folks considered me a bot :) 1132452669 M * Vudumen Bertl: it happens nowadays too :))) 1132452670 M * Vudumen hi :) 1132452744 M * Bertl Vudumen: really? :) 1132452799 M * Vudumen well i think i could find it in my logs :) 1132452806 M * Vudumen (but they just joke :) 1132452838 M * Cru heh 1132453379 Q * Doener Quit: Leaving 1132453768 M * Bertl hmm, who did report the vroot as module issue? 1132455312 M * bwana stupid question alert 1132455334 M * bwana what patch do i use for 2.6.12 1132455486 M * Bertl hmm, why 2.6.12? 1132455499 M * Bertl but, to answer your question, probably vs2.0 1132455500 M * bwana its the latest kernel 1132455541 M * bwana what kernel would one find the most compatability with the current best verion of vserver 1132455553 M * bwana i'm sure you answered this before but i lost my log 1132455604 M * Bertl hmm, no, latest kernel is 2.6.14.2 IIRC 1132455616 M * Bertl and you can get a pretty solid patch for that too 1132455703 M * bwana my bad 1132455708 M * bwana i need sleep 1132455715 M * bwana so my questhio should have been 1132455724 M * bwana what is the best patch for 2.6.14.2 1132455758 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.14.2-vs2.0.1-rc2.diff 1132455768 M * Bertl this is the latest stable release candidate 1132455773 M * bwana very good 1132455777 M * bwana thank you Bertl 1132455787 M * Bertl you#re welcome! 1132456474 J * sebi ~sebi@Fcc33.f.strato-dslnet.de 1132456513 M * bwana hey, what is #linuxchics 1132456535 M * Bertl welcome sebi! 1132456581 Q * sebi_ Ping timeout: 480 seconds 1132456618 M * Vudumen gn 1132456752 M * Cru bye vudu 1132456978 M * Bertl night Vudumen! 1132458154 J * Aiken ~james@tooax6-045.dialup.optusnet.com.au 1132458255 M * Bertl morning Aiken! 1132458337 M * Aiken hello 1132458407 M * Aiken is it worth updating to the 209 tools if the 208 tools still do what I want? 1132458424 M * Bertl hmm, except for testing purposes, no 1132458462 M * Bertl apropos testing, do you feel like checking out a new patch? 1132458489 M * Aiken kernel/patch versions? 1132458506 M * Aiken I still have 2.1.0-rc7 installed on 3 machines 1132458511 M * Aiken rc6 rather 1132458516 M * Bertl well, actually a fix (for the next kernel release) regarding the vroot device 1132458573 M * Aiken so rc7 + patch? I can try that 1132458588 M * Bertl let me explain the issue first ... somebody figured that 1132458612 M * Bertl the vroot device, which works as a proxy for quota ioctls (but you probably know that) 1132458648 M * Bertl can be compiled as module (which is intentional) but doesn't work (if compiled as module) 1132458700 M * Bertl but it works fine when compiled into the kernel 1132458721 M * Bertl it wasn't too hard to figure why ... and I hopefully fixed it by now 1132458731 M * Bertl but of course, it needs testing ... 1132458765 M * Aiken ok 1132458859 M * Bertl interested? 1132458865 M * Aiken yes 1132458895 M * Bertl k, let me prepare the aptch, you might compile a new kernel if you like, but it should work on the older ones too 1132458924 M * Aiken this CONFIG_BLK_DEV_VROOT? 1132458960 M * Bertl yep, precisely 1132459026 M * Aiken I don't have it set 1132459148 M * Bertl okay, you will have to enable quota and this one 1132459313 M * Bertl you will also need to enable quota on a device (could be a loop device or lvm volume) 1132459356 M * Aiken not setup for lvm at the moment but loop is easy 1132459402 M * Bertl okay, best use an ext2 fs, that can be used with quota easily 1132459456 M * Bertl http://linux-vserver.org/Standard+non-shared+quota 1132459471 M * Bertl that's not perfect, but should give some ideas 1132460043 M * micah Bertl: I've been reading over the 2.4 delta you did for me the other day, and I think I understand that the only thing that was missing was the immutable unlink attribute/capability for all the filesystems 1132460075 M * micah and I am guessing that kept the barrier from working properly 1132460107 M * Bertl well, that and one or two other details, IIRC 1132460120 M * micah what I dont understand is why that stuff didn't exist in the other patch 1132460278 M * Bertl probably it was lost in the port/patching ... 1132460393 M * micah I thought it would be much more difficult for me to understand 1132460518 M * Cru a few hours ago there was a question about a replacement for vrebootmgr in vs2.0 - is there some mean to allow restart of all services inside a 2.0 vserver? 1132460669 M * Bertl yes, vrebootmgr is obsolete in 2.0 1132460705 M * bwana whats the deal, why hasn't the kernel people put this into the kernel anyway 1132460729 Q * Aiken Read error: Connection reset by peer 1132460758 M * Bertl bwana: what? linux-vserver? 1132460761 M * Cru I know, but the question was what to use instead ;) 1132460781 J * Aiken ~james@tooax6-045.dialup.optusnet.com.au 1132460790 M * Bertl Cru: the tools (util-vserver 0.30.208/9) should handle it quite fine 1132460791 M * bwana sure 1132460812 M * bwana Bertl: yes 1132460833 M * Bertl bwana: the linux core team does not even accept _obvious_ fixes, and I do not want to bother too much with useless adaptations 1132460842 M * Cru Bertl: so the usual, non-modified sysvinit can control reboot? 1132460843 N * lilo_ lilo 1132460852 M * Aiken Bertl what was that url again please 1132460858 M * Cru have not tried, but will at once... ;) 1132460864 M * Bertl Aiken: http://linux-vserver.org/Standard+non-shared+quota 1132460866 M * Aiken it is 29C in here and just had my 2nd crash in 2 days :( 1132460873 M * Aiken thanks 1132460879 M * Bertl patch will be available in a few minutes 1132460900 M * bwana meh 1132460913 M * Cru well, it claims about missing /dev/initctl as I presumed 1132460925 M * Bertl Cru: who? 1132460970 M * Bertl Cru: out of the blue, Id assume you are using sysv init style, not 'plain' yes? 1132461022 M * Cru eh 1132461045 M * Cru what do you mean with "plain"? 1132461063 M * Bertl well, if you use the 'sysv' init style for your guest 1132461079 M * Bertl then the init process (inside the guest) is not started at all 1132461089 M * Bertl instead, the runlevel scripts are executed 1132461097 M * Cru ah, okay, will check that 1132461103 M * Bertl (this is not how a 'normal' linux server is started) 1132461128 M * Cru okay, I like the old-fashioned way ;) 1132461132 M * Bertl but except for the detail, taht you have to use 'reboot -f' to bypass the missing init, it's very convenient 1132461189 M * Cru ah 1132461223 M * Cru so to allow the vserver to "cleanly" reboot itself I have got to use "plain" init style? 1132461295 M * Bertl no, but to have the 'typical' reboot command work, yes 1132461308 M * Bertl because the typical rebot command does not reboot a guest 1132461315 M * Cru that was what I meant with "cleanly" ;) 1132461326 M * Bertl it just talks to init, and asks 'please reboot the machine' 1132461340 M * Bertl as init is not running in a guest with 'sysv' style 1132461347 M * Bertl this approach fails 1132461358 M * Cru okay, I have no init style configuration for my v2.0 test vservers atm, so sysv is used by default? 1132461388 M * Bertl IIRC, the tools automatically assign one ... 1132461416 M * Cru I think, alias reboot="reboot -f" does the job, too ;) 1132461438 M * Bertl well, yes, that will call into the kernel with a reboot request 1132461449 M * Bertl (which will be handled properly) 1132461609 M * Cru what file is the init style defined in in vserver's configuration dir? 1132461647 M * Cru I did not find anything in the configuration.html supplied with util-vserver 0.30.209... 1132461684 M * Cru neither in the online verison 1132461734 M * Cru ah, apps/init/style 1132461747 M * Cru but no definition on what values are possible 1132461769 M * Bertl they are explained somewhere ... 1132461775 M * Cru also the file was not created by vserver foo build 1132461791 M * Cru at least it is still missing 1132461792 M * Bertl but basically only 'plain' and 'sysv' are used nowadays 1132461806 M * Bertl Cru: it is created if you specify --initstyle 1132461898 M * Cru okay, I did not - but I think it would be nice to have the default value in .defaults/ 1132462048 M * Cru step by step I discover all the great advantages for 2.0 over 1.2... ;) 1132463025 M * Cru hmm, I could create a rpm package providing /etc/profile.d/vserver-sysv.{csh,sh} files that create aliases for reboot and poweroff appending -f flag... 1132463054 M * Cru Bertl: you are using primarily Mandriva, dont you? 1132463091 M * Bertl yes, well, an adapted version 1132463100 M * Bertl and a lot of debian too ... 1132463182 M * Cru well, I know Debian does not have sth like /etc/profile.d/, but does Mandriva use the same format as RedHat/Fedora/CentOS/etc. do? 1132463238 M * Cru sh shells source all /etc/profile.d/*.sh files, csh shells source all /etc/profile.d/*.csh files 1132463278 M * Cru that is making it quite easy to place such shell "extensions" 1132463357 M * Bertl yes, profile.d is very common nowadays 1132463378 M * Cru tell that to the Debian guys ;) 1132463389 J * dddd44 dhb55@218.111.178.108 1132463399 M * Bertl well, debian 'stable' is more than 2 years old, no? 1132463409 M * Bertl dddd44: ping ping ping ! 1132463446 M * Cru well, Debian is a whole book for itself ;) 1132463525 M * Cru I am currently moving almost all woody (v|)servers to CentOS, sarge will follow when unstable enters testing 1132463617 M * Bertl Aiken: finally http://vserver.13thfloor.at/Experimental/delta-vroot-fix01.diff 1132463679 M * Aiken got it 1132463852 M * Bertl compile tests with the following setups would be appreciated too: 1132463858 M * Bertl - vroot disabled 1132463862 M * Bertl - quota disabled 1132463872 M * Bertl - vroot as module 1132463883 M * Bertl - vroot in kernel 1132464382 M * Aiken has the config menu changed between 2.6.14 + vs2.1.0-rc6 and 2.6.14.2+vs2.1.0-rc7? 1132464382 M * Aiken no vroot option under block devices anymore, using the same .config 1132464444 M * Aiken it is there this time 1132464564 M * Bertl yep 1132464573 M * Bertl it now depends on QUOTACTL 1132464610 M * Aiken explains why I got it back when I copied my old config back in a 2nd time 1132465039 M * Bertl heh, sorry forgot to tell ya 1132466164 Q * infowolfe Quit: infowolfe 1132467108 M * Cru Bertl: thanks for info and a relaxed sunday ;) 1132467115 M * Cru bbl 1132467138 Q * Cru Quit: use Unices; $live->free() || die; 1132467171 Q * dddd44 Ping timeout: 480 seconds 1132467285 M * Aiken the no vroot case booted ok 1132467318 M * Bertl good! 1132467545 J * dddd44 dhb55@218.111.178.108 1132467957 M * Aiken with the no vroot and no quota, what am I looking for? 1132467974 M * Aiken just the machine booting, everything running as per normal, no errors and so on? 1132467983 M * Bertl just normal compile and bootup 1132467995 M * Bertl no further tests required 1132470275 M * Aiken no bad messages or crashes using vroot as a module 1132470293 M * Bertl excellent, does the quotaioctl work too? 1132470698 M * Aiken just searching for a working pastebin 1132470712 M * Bertl pastebin.com broken? 1132470730 M * Aiken for me it is 1132470739 M * Aiken http://pastebin.ca/29509 1132471441 M * Bertl hmm, and /dev/hdv1 is a configured vroot device? 1132471495 M * Aiken yes 1132471520 M * Bertl hmm, could you try with: 1132471538 M * Bertl 'quotaon -Fvfsv0 -aug 1132471578 M * Bertl and if that fails too, does it work with the 'real' device copied to /dev/hdv1? 1132471860 M * Bertl okay, I'm pretty tired right now, found a minor issue which might crash the kernel if the vroot device is used before the vroot driver was loaded 1132471892 M * Bertl change is in fs/quota.c ~line 403 1132471913 M * Bertl if (bdev && bdev->bd_inode && 1132471913 M * Bertl + vroot_get_real_bdev && 1132471913 M * Bertl imajor(bdev->bd_inode) == VROOT_MAJOR) { 1132471941 M * Bertl (but should have no effect on your current results) 1132471991 M * Aiken I'll keep playing 1132471995 M * Bertl Aiken: thanks a lot for testing so far ... 1132472030 M * Bertl just let the channel know about the results ... I will read up later 1132472049 M * Bertl good night everyone .. cya tomorrow! 1132472058 N * Bertl Bertl_zZ 1132478200 Q * Aiken Ping timeout: 480 seconds 1132482691 J * kridian ~kridian@196.211.0.69 1132485841 J * lilo_ ~lilo@lilo.usercloak.oftc.net 1132485942 Q * lilo Ping timeout: 480 seconds 1132486160 J * renihs ~renihs___@193.170.52.70 1132486898 Q * renihs Quit: Leaving 1132488711 Q * kridian Quit: 1132489719 M * sizo moin 1132490579 J * shedi ~siggi@inferno.lhi.is 1132496080 J * Larsioa ~lars@you.are.the.lamer.la 1132496081 M * Larsioa Hello 1132496089 M * Larsioa does anybody know if its possibel to get oidentd to listen on a ip 1132496093 M * Larsioa and not 0.0.0.0 1132496363 M * Larsioa :( 1132496381 M * Larsioa Nov 20 22:11:32 esben oidentd[32399]: Fatal: Unable to setup listening socket 1132496519 M * ag- Larsioa: hacking the source code 1132496670 M * Larsioa hm heh 1132496671 M * Larsioa ;/ 1132496748 M * Larsioa -a or --address= 1132496748 M * Larsioa Listen for connections on the specified address. The default is to listen for connections on all configured IP addresses. 1132496751 J * Viper0482 ~Viper0482@p54974A0A.dip.t-dialin.net 1132496768 M * Larsioa but that doesnt work 1132496861 M * Larsioa ah worked now. 1132497019 J * Doener doener@i5387F50F.versanet.de 1132497079 Q * Viper0482 Remote host closed the connection 1132501445 Q * Doener Ping timeout: 480 seconds 1132501458 J * Doener doener@i5387F33C.versanet.de 1132502166 J * bambr ~shurik-a3@birulevonet-gw.fiord.ru 1132502166 N * Bertl_zZ Bertl 1132502173 M * Bertl morning folks! 1132502173 P * bambr 1132502240 M * michal_ yy.. "morning" Bertl :) 1132503209 M * derjohn michal_, , mind the austrian timezone ;) 1132503240 M * Larsioa :D 1132503365 M * michal_ 17:15 here, not really different :) 1132503390 M * michal_ it is about the same ;] 1132503543 M * Larsioa same :D 1132503547 M * Larsioa <- Denmark 1132503913 M * Doener derjohn: you mean the BPT ;) 1132503920 M * derjohn s/ austrian timezone/berl timezone/ 1132503921 M * derjohn :) 1132503922 M * Doener (Bertl-Personal-Timezone) 1132503931 M * Larsioa hah :D 1132503958 M * derjohn a subclass of geek-tz, which I am fully compatible ;) 1132503987 M * Bertl excellent :) 1132504037 M * derjohn Larsioa, did yiu check out ispman? 1132504100 M * Larsioa derjohn: no im buying a Sun cobalt RAQ 550 cheap with webhosting panel. 1132504139 M * derjohn Larsioa, *ouch* 1132504148 M * Larsioa thoose sun cobalt are sooo sweet :P 1132504152 M * Larsioa cool design and cool LCD :P 1132504170 M * derjohn which won't work with debian unless you do natsy things 1132504174 M * derjohn (last time i checked) 1132504187 M * Larsioa i can be nasty. 1132504223 M * derjohn our's was called "kitt", because the panel acts like a "knight 2000 font led row" 1132504242 M * Larsioa coool :P 1132504253 M * Larsioa but what does the server come with standard? 1132504259 M * derjohn maybe "Zylone" would fit, too. 1132504274 M * derjohn Modified redhat ... 1132504280 M * Larsioa Okay. 1132504289 M * derjohn with a user called "admin" 1132504300 M * Larsioa Do you know if there panel works under Debian too? 1132504301 M * derjohn I used RAQ4 though 1132504303 M * Larsioa the webhosting panel. 1132504321 M * derjohn but there is a sf.net project for reenigineer the rom etc. 1132504327 M * Larsioa Well i dont know how big changes there is in RAQ5 but there are 2 harddisks. 1132504328 M * Bertl okay, off for dinner, back shortly ... 1132504334 N * Bertl Bertl_oO 1132504349 M * derjohn Larsioa, think about it TWICE, 1132504360 M * Larsioa ? 1132504382 M * derjohn for linux-vserver you will need a new kernel anyway and the cobalt OS is opensource meanwhile 1132504410 M * derjohn vhcs and syscp install like ... every noob can do it. 1132504410 M * Larsioa well ill not run vserver on the cobalt. 1132504413 M * Larsioa only webhosting. 1132504435 M * derjohn as you like. dont forget to buy a second one as backup. 1132504445 M * Larsioa ill just create a RAID 1132504449 M * derjohn the os shipped will only run on cobalt HW AFAIR 1132504462 J * Alissa ali2@151.81.1.103 1132504465 M * derjohn super! if the mobo dies 1132504471 M * Larsioa http://www.raqdevil.com/ 1132504471 M * Alissa hi all 1132504472 M * Larsioa seems sweet. 1132504484 M * derjohn Larsioa, HF with it ;) 1132504501 M * Larsioa The RAQdevil is software that provides web based management of multiple aspects of a server. RAQdevil manages system users, web sites, email, DNS, and other common server components used for web hosting. 1132504502 M * Larsioa ;) 1132504508 M * derjohn there is pkgmaster.? for getting curremt versions 1132504519 M * derjohn k 1132504523 M * Larsioa derjohn: well i think it would be more secure than the Sun OS webmanagement system 1132504530 M * Larsioa as i remember there is some security bugs 1132504534 M * Larsioa and its not updated anymrope. 1132504536 M * Larsioa anymore* 1132504550 M * derjohn ay caramba .... 1132504570 M * Larsioa ? 1132504608 M * derjohn Larsioa, I cannot understand that you go for FOSS here and then ... you buy something proprietary for webhosting? 1132504613 Q * aba Remote host closed the connection 1132504625 M * Larsioa derjohn: FOSS? 1132504631 M * Larsioa and whats wrong with Sun cobalt? 1132504641 M * derjohn Free Open Soure Software = FOSS 1132504678 M * derjohn see above :) 1132504688 M * Larsioa Sun Cobalt RaQ550 software. Sun released the code for the RaQ under a BSD-style license in December of 2003 and the code was then ported to Fedora Linux by the BlueQuartz project. 1132504700 M * derjohn But I dont say anything bas about it 1132504719 J * aba ~aba@eos.turmzimmer.net 1132504754 M * Larsioa derjohn: The sun cobalt machine would be fine with FreeBSD on it and the RAQDevil 1132504757 M * derjohn ah, why not a speedy HW mit this bluefarz? 1132504771 M * derjohn ok 1132504777 M * derjohn :) 1132504792 M * Larsioa :) 1132504877 M * Larsioa http://www.raqdevil.com/screenshots.html 1132504879 M * Larsioa ups 1132504886 M * Larsioa http://www.raqdevil.com/install.html and the installation seems soo easy 1132505113 M * derjohn e.g. http://www.syscp.de/docs/installation/debian/sarge/basicinstallation/english 1132505150 M * derjohn look it could be done by a "mechanicalö turk" (as google would call) 1132505159 M * derjohn though I go for ispman 1132505258 M * Larsioa :) 1132505262 M * Larsioa RAQdevil looks hot. 1132505386 M * derjohn ok ok .... I understood that you are an apprentice of satan^w the devil :))) 1132505439 M * Larsioa :P :P :P 1132505488 M * Larsioa But hmm does thoose systems has userbased control panel to there websites created by the admin? 1132505509 M * Larsioa does the default Raq4 has that derjohn ? 1132505523 M * Larsioa just got a yest from the raqdevil crew cool then. 1132505569 M * derjohn it runs colbalt os, installed as image via netboot 1132506115 N * Bertl_oO Bertl 1132506119 M * Bertl back now ... 1132506126 M * Larsioa sweet love. 1132507585 Q * aba Remote host closed the connection 1132507764 Q * FireEgl Ping timeout: 480 seconds 1132508087 J * aba ~aba@eos.turmzimmer.net 1132508237 M * Bertl wb aba! 1132508551 Q * shedi Quit: Leaving 1132508578 J * FireEgl Atlantica@2001:5c0:84dc:: 1132509958 M * AndrewLee hi 1132509963 M * Bertl hey AndrewLee! 1132510054 M * AndrewLee Bertl: I followed the FAQ G.2(Is DHCP possible in a vserver?) to set bcast to 255.255.255.255, and then dhcp still doesn't work. 1132510077 M * AndrewLee Bertl: I got this error message in syslog: 1132510078 M * AndrewLee Nov 21 02:01:02 ping dhcpd: Open a socket for LPF: Operation not permitted 1132510137 M * AndrewLee Bertl: I think it might need some settings in bcapabilities file, am I right? 1132510180 M * Bertl yes, most likely the broadcast capability 1132510203 M * Bertl but it would be good to see an strace -fF of the dhcp 1132510207 A * AndrewLee reading man capabilities 1132510255 M * Bertl grep -i broadcast /usr/include/linux/capability.h 1132510255 M * Bertl /* Allow broadcasting, listen to multicast */ 1132510255 M * Bertl #define CAP_NET_BROADCAST 11 1132510416 M * AndrewLee I found this error in strace: 1132510417 M * AndrewLee ioctl(7, SIOCGIFHWADDR, {ifr_name="eth1", ifr_hwaddr=00:d0:b7:b2:67:0a}) = 0 1132510417 M * AndrewLee socket(PF_PACKET, SOCK_PACKET, 768) = -1 EPERM (Operation not permitted) 1132510438 M * Bertl that looks like CAP_NET_RAW 1132510476 M * Alissa Bertl hi :) 1132510481 M * Bertl hey Alissa! 1132510512 M * Alissa Bertl pach vserver on the kernel 2.6 is stable? 1132510513 M * AndrewLee Bertl: I found capabilities's manpage says CAP_NET_BROADCAST (Unused) 1132510546 M * AndrewLee Bertl: Let me try both CAP_NET_BROADCAST and CAP_NET_RAW to see which one is required 1132510559 M * Bertl well, let me check for the broadcast cap :) 1132510650 M * Bertl yep, indeed, that capability is not used anylonger ... 1132510687 M * AndrewLee Bertl: So, souldn't I try it a try? 1132510745 M * Bertl yes, try with the NET_RAW one ... but the BROADCAST has no effect 1132510779 M * Bertl but it looks like we could use the broadcast in the future ... 1132510784 M * Bertl thanks for the information 1132510809 M * AndrewLee Bertl: What idea you got for using the broadcast? 1132510880 M * AndrewLee Bertl: Yes, CAP_NET_RAW works. Should I update that in FAQ? 1132510924 M * AndrewLee Bertl: The answer in FAQ is still using the legancy config, should I update it to use new-style config? 1132511660 M * Bertl well, I think it might be an 'option' to use cap_broadcast for enabling/disaling the broadcast IPs for a given network 1132511672 M * Bertl and yes, please update the FAQ 1132511860 Q * Alissa Quit: Hasta La Dominación, siempre... 1132512067 M * AndrewLee Bertl: Looks like it's unnecessary to add 255.255.255.255 into bcast file. 1132512109 M * AndrewLee Bertl: After added CAP_NET_RAW, dhcp works itself.(I tested on dhcp3-server) 1132512322 M * Bertl okay, maybe make that 'might be unnecessary' 1132512342 M * Bertl because we can not tell for all dhcp servers 1132512480 M * AndrewLee Bertl: Thanks. 1132512506 M * Bertl thanks to you for testing! 1132513102 M * AndrewLee Bertl: I added the solution to FAQ page, please have a loook, I am not sure that's correct or not. And thank you for your help. :) 1132513347 J * Aiken ~james@tooax6-026.dialup.optusnet.com.au 1132513560 M * Bertl AndrewLee: you're welcome! will do so! 1132513567 M * Bertl morning Aiken! 1132513586 M * Aiken hi 1132513596 M * Aiken I got no where yesterday 1132513613 M * Bertl np, we will add a bunch of 'debug' messages today 1132513631 M * Bertl Iw anted to do that for a long time now, but it just didn't happen yet :) 1132513664 M * Aiken I wouldn't mind a bit of time to wake up properly, couldn't sleep and sun rise was only 17 min ago 1132513704 M * Bertl take as long as you like ... 1132514190 Q * Hunger Remote host closed the connection 1132514276 J * Hunger Hunger.hu@Hunger.hu 1132514631 M * Bertl wb Hunger! 1132514845 Q * sladen Ping timeout: 480 seconds 1132515176 J * sladen paul@starsky.19inch.net 1132516541 M * Aiken Bertl I think the 1st thing to do would be to make sure I was not doing anything silly yesterday 1132516612 M * Bertl I guess the debug output will help there, will take a few minutes to add it 1132516660 M * Aiken with vrsetup which device should I associate vroot0 with? 1132516684 M * Aiken I tried the device the vserver filesystem is on 1132516698 M * Aiken also tried loop0 pointing to a file that was formatted ext2 1132516735 M * Bertl basically vroot points to the real device (which is the rootfs of the guest) 1132516742 M * ag- is it possible to have more than 8 vroot devices? it requires patching, i suppose? 1132516769 M * Bertl nope, just an option (upper limit is 256) 1132516778 M * Aiken /dev/etherd/e9.0 on /vservers type ext3 (rw,data=ordered) 1132516797 M * Aiken I did try vrsetup /dev/vroot0 /dev/etherd/e9.0 1132516813 M * Bertl yes, that looks good 1132516885 M * Aiken (root@pebbles) cat /etc/vservers/hoppy/apps/init/mtab 1132516885 M * Aiken /dev/hdv1 / ufs rw,usrquota,grpquota 0 0 1132516905 M * Bertl looks good too, filesystem on e9.0 is? 1132516908 M * Aiken (root@pebbles) cat /etc/vservers/hoppy/ccapabilities 1132516908 M * Aiken quota_ctl 1132516910 M * Aiken ext3 1132516934 M * Aiken so easy enough to mount ext2 if needed 1132516937 M * ag- my question was quite stupid, i'm gonna torture-test it anyway :) 1132516962 M * Bertl max_vroot=16 will give you 16 vroot device 1132517992 M * Aiken I mounted /vservers ext2 and no error messages :) 1132518012 M * Bertl hmm ... 1132518213 M * Aiken last thing I want now is a crash, I put ext2 up there with fat for reliabilty with crashes and power failures 1132518449 M * Bertl so, it works with ext2, but fails with ext3, yes? 1132518475 M * Aiken yes 1132518495 M * Bertl hmm, did you mount the ext3 with any quota options? 1132518510 M * Bertl does the ext3 have quota enabled in the kernel config? 1132518532 M * Larsioa y0 1132518568 M * Aiken in filesystems there is no mention of quota for ext2 or ext3 1132518575 M * Aiken just the quota options further down the page 1132518634 M * Bertl hmm, okay, let's check if it works if you mount the ext3 with usrquota or 'jsut' quota 1132519059 M * Aiken /dev/etherd/e9.0 on /vservers type ext3 (rw,data=ordered,usrquota) 1132519074 M * Aiken that seems to work ok, no errors or warnings 1132519100 M * Bertl okay, excellent! 1132519124 M * Aiken that was with vroot compiled into the kernel 1132519130 M * Aiken about to try it as a modules 1132519261 Q * mnemoc Ping timeout: 480 seconds 1132519404 M * Hollow Bertl: yay, util-vserver works with diet :) 1132519419 M * Bertl hey great! url? binaries? 1132519424 M * Hollow sec 1132519470 M * matti ;D 1132519491 M * micah Bertl: what is the purpose of the name changes in delta-2.4.27-9-vs1.2.10-fix01.diff (changing IS_IMMUTABLE_FILE with IS_IMMUTABLE)? 1132519509 M * Bertl micah: the change has two reasons actually ... 1132519528 M * Bertl - we wanted to leave the 'default' as *_IMMUTABLE 1132519533 M * Hollow Bertl: http://dev.gentoo.org/~hollow/vserver/libvserver/libvserver-1.0_pre20051119.tar.bz2 and http://dev.gentoo.org/~hollow/vserver/util-vserver/util-vserver-1.0_pre20051120.tar.bz2 1132519563 M * Bertl micah: and we added the XOR 1132519577 M * micah (IS_IMMUTABLE_LINK replaces IS_IXORUNLINK) 1132519585 M * micah (IS_IMMUTABLE_LINK is replaced by IS_IXORUNLINK) 1132519589 M * micah second one :) 1132519634 M * Bertl yes 1132519646 M * Hollow i had to reimplement the iattr tool 3 times, becuase diet doesn't know fts nor nftw, and ftw is somewhat useless 1132519662 M * Hollow i'm wondering if uclibc is more glibc compliant 1132519702 M * micah Bertl: ok, leave 'default' as *_IMMUTABLE and add the XOR are the two reasons, but I guess I do not understand the purpose of that 1132519707 M * Bertl had a look at uclibc some time ago, and it was not that bad (on x86) 1132519727 M * Bertl micah: let me give you an example: 1132519736 M * Hollow i'll probably set up a cross-gcc chroot, to test uclibc with util-vserver 1132519741 M * micah (because I dont understand the kernel code that well, not because I question you :) 1132519817 M * Bertl micah: consider out of kernel code, which, for whatever reason, uses IS_IMMUTABLE() 1132519859 M * Hollow Bertl: we have to discuss the reboot issue again imo, also the mentioned netlink/whatever state tracker would be handy with this 1132519867 M * Hollow (i.e. reboot from inside) 1132519873 M * Aiken it works with vroot compiled in 1132519875 M * Bertl micah: with this change, it will work as expected 1132519888 M * Aiken but with vroot as a modules I am back to 1132519889 M * Aiken quotaon /dev/hdv1 1132519890 M * Aiken quotaon: using //aquota.group on /dev/hdv1 [/]: No such device 1132519890 M * Aiken quotaon: using //aquota.user on /dev/hdv1 [/]: No such device 1132519892 J * mnemoc ~amery@200.75.27.59 1132520031 M * micah Bertl: maybe I should rephrase my question: why would it would be better to use fix01 with the name changes? 1132520056 M * Bertl ah, okay, that is easier to explain :) 1132520067 M * Bertl micah: again for two reasons: 1132520090 M * Bertl micah: first, because the code _with_ the anme changes is easier to read 1132520122 M * Bertl micah: and second, if we do a delta (e.g. for 1.2.11 or later) it will apply ontop of that 1132520241 M * micah that makes a lot of sense 1132520330 M * micah I have to write a high-level summary of the issue so that it can be updated in sarge 1132520425 M * micah Would it be correct to say: "The cause of this is the barrier code included in the Debian patch improperly handles immutable unlink filesystem attributes and the necessary capabilities to protect against it. When the port was done to Debian key pieces were not included that would prevent such an escape. Namely the immutable unlink extended filesystem attributes and the capability system to handle those." 1132520532 M * micah that is a little redundant, cleaning that up a little 1132520561 M * micah but I just wanted to make sure I was not misrepresenting the issue  1132520576 M * Bertl sounds good to me 1132520582 M * micah btw. i am making it clear that this is a debian specific patch problem, not something from linux-vserver 1132520692 Q * mejlholm Quit: Client exiting 1132520839 M * Larsioa derjohn: syscp is cute. 1132520865 M * Larsioa much better than vhcs 1132520905 M * micah Bertl: if I have patch-2.4.27-9-vs1.2.10-2.diff is there an easy way i can apply delta-2.4.27-9-vs1.2.10-fix01.diff to that patch? 1132520981 M * micah patching a patch is confusing 1132520985 M * derjohn Larsioa, vhcs.org, top 1132521005 M * derjohn s/top/too :) 1132521043 M * derjohn Larsioa, I just installing an ISPMAN at the time with a Makefile i wrote last year. seems to work still. 1132521058 M * derjohn 5 times return and the thing is pretty much uasable ;) 1132521080 M * Larsioa ;) i think syscp is nice its much more simpler than vhcs and syscp has a nice simple design and easy to use. 1132521082 M * derjohn btw: is's ISPMAN.net (not .org as I told) 1132521093 M * derjohn Larsioa, yes 1132521107 M * derjohn for a one-box-setup this is straoghtforward 1132521111 M * Larsioa derjohn: looking at it now. Has ispman userinterface for the customers too? 1132521126 M * derjohn Larsioa, yes, provisioning, too 1132521138 M * derjohn i.e. you can set limits for max domain per reseller etc. 1132521145 M * derjohn vhcs has that feature too. 1132521157 M * Larsioa ye but vhcs has to much fancy design. 1132521159 M * Larsioa too heavy 1132521162 M * Larsioa dont like it. 1132521168 M * Bertl micah: just add the delta ... 1132521169 M * derjohn there is also gnuhh (hosting helper), wich has also a webinterface for filewall stuff per domain 1132521186 M * Bertl micah: i.e. cat patch-2.4.27-9-vs1.2.10-2.diff delta-2.4.27-9-vs1.2.10-fix01.diff >patch-2.4.27-9-vs1.2.10-3.diff 1132521197 M * derjohn Larsioa, syscp is very German-driven AFAIR :) 1132521212 M * Larsioa derjohn: there is a english template :) 1132521219 M * derjohn Larsioa, "deutsche Wertarbeit" (difficult to translate) 1132521225 M * derjohn Larsioa, :) 1132521242 M * Larsioa ;) 1132521316 M * Larsioa wow derjohn 1132521321 M * Larsioa i found an nice one. 1132521327 M * Larsioa http://www.ispconfig.org/ 1132521332 M * Larsioa look at the cool smooth design 1132521421 M * micah Bertl: hrm, i get one hunk failure 1132521530 M * derjohn micah, on Debian? 1132521577 M * derjohn micah, or in the bertlsch' incremental use? 1132521603 M * Larsioa derjohn: saw it? :) 1132521683 M * Bertl micah: well, the patch was against your sources IIRC 1132521702 M * Bertl micah: didn't you test it before? 1132521720 M * micah yes, I see that the original patch modifies xfs_vnode.c as does the delta 1132521755 M * Bertl that should be fine ... 1132521775 M * Bertl maybe you are 'testing' with --dry-run? 1132521798 M * derjohn Larsioa, no that is new. We should write a paper abiut existing free hosting solutuions? 1132521839 M * Larsioa derjohn: yes! a good idea. 1132521941 M * micah Bertl: no, I think it is just a spacing issue 1132521957 M * Bertl then try applying with -l 1132521968 M * Bertl and after that, rebuild the patch with: 1132521969 M * derjohn Larsioa, you take the cobalt part :) 1132521992 M * Bertl micah: diff -NurpP --minimal kernel-without-patch kernel-with-patch 1132522089 M * Larsioa derjohn: suure, ill get my cobalt this week :P 1132522142 M * micah Bertl: kernel-with-patch = kernel-with-patch-2.4.27-9-vs1.2.10-2.diff? 1132522160 M * micah (I think the problem is just an indentation difference) 1132522176 M * Bertl micah: do the following: 1132522187 M * Bertl - first, get the debian base kernel 1132522200 M * Bertl - unpack it (apply the debian patches, etc,etc) 1132522210 M * Bertl - make a shallow copy like this: 1132522243 M * Bertl cp -la linux-2.4.27-9 linux-2.4.27-9-vs1.2.10-new 1132522262 M * Bertl - apply the 'old' vserver patch like this: 1132522288 M * Bertl (cd linux-2.4.27-9-vs1.2.10-new; patch -p1 -l <../patch-2.4.27-9-vs1.2.10-2.diff ) 1132522300 M * Bertl - apply the 'delta' like this: 1132522320 M * Bertl (cd linux-2.4.27-9-vs1.2.10-new; patch -p1 -l <../delta-2.4.27-9-vs1.2.10-fix01.diff) 1132522336 M * Bertl - create a new patch like this: 1132522347 M * Bertl diff -NurpP --minimal linux-2.4.27-9 linux-2.4.27-9-vs1.2.10-new 1132522349 J * shedi ~siggi@inferno.lhi.is 1132522434 M * micah ok 1132522440 M * micah that all went well 1132522472 M * micah now i should be able to create another shallow copy of the original debian kernel source and apply the new patch? 1132522504 M * Bertl yep 1132522561 M * micah I think the ignore whitespace -l option did it 1132522587 M * Bertl very likely, but you can 'retry' the entire procedure without 1132522599 M * micah I will do so 1132522606 M * Bertl (with a new kernel name) 1132522611 M * micah of course :) 1132522663 M * micah yes, if I do it again without -l it fails 1132522899 M * Bertl k, strange but so be it :) 1132522950 M * micah theoretically I should be able to interdiff the new patch with the old patch and get the delta fix01 1132522968 M * Bertl theoretically yes :) 1132522986 M * micah but interdiff is not so smart 1132523013 M * Bertl well, first, very different looking patches can give the same changes 1132523057 M * micah yeah, interdiff thinks that the new patch removes the Configure.help info for CONFIG_BLK_DEV_VROOT, but just a simple look at the new patch shows that does not happen 1132523068 M * Bertl an second, not all changes can be detected 1132523546 Q * Larsioa Read error: Connection reset by peer 1132523871 Q * shedi Quit: Leaving 1132523885 M * Bertl Aiken: http://vserver.13thfloor.at/Experimental/delta-vroot-clean01.diff 1132523893 M * Bertl Aiken: not that it is needed anymore :) 1132523918 Q * Johnnie Quit: G'bye! 1132524025 J * Johnnie ~john@acs-24-154-53-217.zoominternet.net 1132524031 Q * Aiken Quit: Leaving 1132524102 M * ag- Hollow: does the new tools version still works fine with glibc? 1132524299 M * Hollow i think so, yes 1132524450 Q * Johnnie Quit: G'bye! 1132524474 J * Johnnie ~john@acs-24-154-53-217.zoominternet.net 1132524480 Q * Johnnie Quit: 1132524579 J * Johnnie ~john@acs-24-154-53-217.zoominternet.net 1132524847 Q * Johnnie Quit: 1132525173 J * Johnnie ~john@acs-24-154-53-217.zoominternet.net 1132525188 M * Bertl Johnnie: ?? 1132525203 M * Johnnie Sorry, E17 was being funny. 1132525210 M * Johnnie CVS snapshots... blah. 1132525213 M * Johnnie It's all better now. 1132525217 M * Bertl k 1132525621 M * Hollow ok, off to bed now, Bertl: did you already give the tools a (compilation) try? 1132525646 M * Bertl no, sorry, forgot about that ... 1132525653 M * Hollow no problem, cu tomorrow then 1132525709 M * Bertl good night! will test them now ... 1132525713 J * Aiken ~james@tooax6-238.dialup.optusnet.com.au 1132525736 M * Hollow ok, i may be watching some more minutes, just tell me :) 1132525844 M * Bertl comments: 1132525873 M * Bertl - would it be possible to give a config overview (to see the pathes after configure)? 1132525874 Q * dddd44 Read error: Connection reset by peer 1132525925 M * Bertl Hollow: wow, make bails out with a bunch of errors 1132525942 M * Bertl In file included from nidof.c:28: 1132525942 M * Bertl /usr/include/vserver.h:250: error: parse error before "xid_t" 1132525942 M * Bertl /usr/include/vserver.h:275: error: parse error before "vc_new_s_context" 1132525957 M * Hollow did you install libvserver before? 1132525966 M * Hollow (and probably remove an old util-vserver install) 1132526005 M * Bertl hmm, could you walk me through the install process? 1132526023 M * Bertl I mean, the tools should not depend on stuff, right? 1132526030 M * Hollow remove util-vserver, install libvserver, install util-vsevrer 1132526046 M * Hollow i should probably add a check for libvserver to be the right one 1132526086 M * Bertl okay, util-vserver deinstalled 1132526100 M * Bertl rerunning configure 1132526114 M * Bertl no complaints there 1132526133 M * Bertl make fails now with looking for the lib ... 1132526139 M * Bertl (a check would be nice :) 1132526148 M * Hollow yeah 1132526155 M * Bertl where do I get the latest libvserver? 1132526165 M * Hollow http://dev.gentoo.org/~hollow/vserver/libvserver/libvserver-1.0_pre20051119.tar.bz2 1132526201 M * Bertl ah, already have it, tx 1132526249 M * Bertl libvserver fails here: 1132526254 M * Bertl bcaps-list.c: In function `bcaps_list_init': 1132526254 M * Bertl bcaps-list.c:76: error: `CAP_AUDIT_WRITE' undeclared (first use in this function) 1132526279 M * Bertl which is not unexpected, as the cap is pretty new, and the kernel headers are older 1132526279 M * Hollow mhm, which kernel header version? 1132526284 M * Hollow k 1132526294 M * Hollow go to bcaps-list.c and remove the entry 1132526328 M * Bertl done 1132526345 M * Bertl now it fails here: 1132526347 M * Bertl personalityflag-list.c: In function `persflag_list_init': 1132526347 M * Bertl personalityflag-list.c:46: error: `FDPIC_FUNCPTRS' undeclared (first use in this function) 1132526355 M * Hollow same here i assume 1132526358 M * Bertl personalityflag-list.c:48: error: `ADDR_COMPAT_LAYOUT' undeclared (first use in this function) 1132526361 M * Bertl personalityflag-list.c:49: error: `READ_IMPLIES_EXEC' undeclared (first use in this function) 1132526364 M * Bertl personalityflag-list.c:54: error: `ADDR_LIMIT_3GB' undeclared (first use in this function) 1132526404 M * Hollow ok, so add a check for kernel headers as well 1132526418 M * Bertl would be cool to either make those defines optional, or to add them to your sources 1132526438 M * Hollow yeah 1132526457 M * Bertl personality-list.c:48: error: `PER_LINUX_FDPIC' undeclared (first use in this function) 1132526461 M * Bertl personality-list.c:59: error: `PER_LINUX32_3GB' undeclared (first use in this function) 1132526464 M * Bertl personality-list.c:66: error: `PER_OSF4' undeclared (first use in this function) 1132526467 M * Bertl personality-list.c:67: error: `PER_HPUX' undeclared (first use in this function) 1132526518 M * Bertl k, compiled now ... 1132526837 M * Hollow util-vserver ready too? 1132526874 M * Bertl yep, sec, just extracting names for a binary package (for testing in QEMU) 1132526884 M * Hollow ok 1132526903 M * Hollow to test a guest, you need to create /etc/vservers/name/context.conf with at least VX_XID="123" and VX_INIT="plain" 1132527095 Q * sladen Ping timeout: 480 seconds 1132527218 J * sladen paul@starsky.19inch.net 1132527798 M * daniel_hozac single file configuration again? :| 1132527884 M * Bertl Hollow: still around? 1132527912 M * Hollow yeah 1132527936 M * Bertl okay, installed the tools now .. do you have a modified testme.sh? 1132527958 M * Bertl or should it work out of the box? 1132527973 M * Hollow no, but you can just give it a try 1132528008 M * Bertl okay, fails ... do you plan to support long options too? 1132528030 M * Hollow yeah, i could do that 1132528041 M * Bertl how does one detect your tools/the version? 1132528089 M * Hollow every tool supports -V (version switch) 1132528101 M * Hollow it should show 1.0 1132528114 M * Bertl ah, because it's not listed on the help :) 1132528164 M * Hollow which command failed? 1132528175 M * Bertl /usr/sbin/vcontext: invalid option -- - 1132528187 M * Bertl basically the first one :) 1132528198 M * Hollow if you set UVDEBUG it will show some bash debugging 1132528230 M * Bertl well, it's obvious, it's the double dash 1132528267 M * Bertl or? 1132528311 M * Hollow hm, the thing is, the vserver script does not pass -- to vcontext 1132528392 M * Bertl chcontext --ctx 1 grep -Ee '[[]|s_context|VxID' /proc/self/status 1132528395 M * Bertl this fails ... 1132528432 M * Bertl but obviosuly it uses enricos chcontext 1132528440 M * Hollow which command did you execute? 1132528440 M * Hollow (the command line) 1132528440 M * Hollow there is no chcontext, this is a leftover of old util-vserver installation probabl 1132528460 M * Bertl yes, but it will fail because of the missing chcontext anyway 1132528478 M * Hollow try: vcontext -M -x 1 -- grep -Ee '[[]|s_context|VxID' /proc/self/status 1132528517 M * Bertl yup, works, okay will try to add tool detection ... 1132528560 M * Hollow could you try the vattr tool quickly? 1132528577 M * Bertl sure .. what test? 1132528604 M * Hollow it's a combined replacement for {ls,ch}xid, {show,set}attr.. e.g.: vattr /proc 1132528619 M * Hollow or: vattr -S -f HIDE /proc/mounts 1132528642 M * Bertl seems to work fine 1132528648 M * Hollow great :) 1132528658 M * Hollow do you have a filesystem with tagxid? 1132528683 M * Bertl sure 1132528698 M * Hollow try to list it, the second column should show the xids 1132528739 M * Bertl how do I change the xid? 1132528761 M * Hollow vattr -S -x 123 /foo/bar 1132528787 M * Bertl vattr -S -x 123 /mnt/part1/ZZZZ 1132528788 M * Bertl Segmentation fault 1132528801 M * Hollow ic 1132528826 M * Hollow does it at least list 00000 for existing files? 1132528840 M * Bertl yes 1132528847 M * Hollow ok, i'll look into that tomorrow 1132528881 M * Bertl k, np ... 1132528917 M * Hollow ok, feel free to test the other apps too, i'm really off to bed then 1132528925 M * Bertl k, good night! 1132529005 M * Bertl okay, looks like I'm off to bed too ... little tired today ... 1132529025 M * daniel_hozac good night! 1132529030 M * Bertl have a good whatever, everyone ... cya 1132529035 N * Bertl Bertl_zZ 1132530193 J * Cru ~mindwarp@instructor.e.de.wahlich.com 1132530201 M * Cru hi