1160265611 Q * _node Ping timeout: 480 seconds 1160266003 M * azazel i'm just looking the effort involved in adding configuration extensions... 1160266011 Q * fs Ping timeout: 480 seconds 1160266135 J * ntrs ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1160266368 M * daniel_hozac hmm? 1160266514 J * fs fs@213.178.77.98 1160266885 M * daniel_hozac what problem is it that you think would be solved by using Python? 1160267011 M * azazel uhumm... creating an /etc//nagios/[alarms|contacs|commands] structure and maybe adding a "vserver whatever maintenance" command and some "vserver create" options... 1160267109 M * daniel_hozac and which of those do you find difficult? 1160267148 M * daniel_hozac should be straight-forward. 1160267266 M * azazel yep, but i have to patch the original util-vserver's scripts code.... 1160267297 M * daniel_hozac well, umm, yeah. 1160267524 M * daniel_hozac note that it's entirely possible to add configuration options without patching. 1160267697 M * micah daniel_hozac: i think the later, i'm hoping I can ask aba for help to get it in if it doesn't make it (since he uses it and is a release manager and would probably be interested in it happening) 1160267770 M * azazel doh... "vserver build" has special handling 1160267833 M * micah daniel_hozac: I could not think of a justification, according to debian policy, for making the urgency medium, although I may have misread 1160267877 M * azazel mmm.... there are problems with debs 1160267878 M * azazel ? 1160268116 M * micah azazel: hmm? 1160268412 M * daniel_hozac micah: ok, cool. 1160269201 N * Bertl_zZ Bertl 1160269215 M * Bertl hey folks! 1160269226 M * azazel hi betl 1160269235 M * azazel ops... sorry... Bertl 1160269243 M * Bertl np :) 1160269507 M * Bertl doener: ping? 1160273668 Q * ensc Killed (NickServ (GHOST command used by ensc_)) 1160273678 J * ensc ~irc-ensc@p54B4DBF0.dip.t-dialin.net 1160274201 M * Bertl okay, guess I'm off to bed again ... have a good one! 1160274210 N * Bertl Bertl_zZ 1160279382 Q * Curus helium.oftc.net strange.oftc.net 1160279430 J * Curus ~Curus@kbhn-vbrg-sr0-vl209-213-185-8-10.perspektivbredband.net 1160280004 J * The_Shadow ~The_Shado@pool-151-205-184-163.ny325.east.verizon.net 1160280055 P * The_Shadow 1160281057 J * s0undt3ch_ ~s0undt3ch@bl7-243-165.dsl.telepac.pt 1160281179 Q * s0undt3ch Ping timeout: 480 seconds 1160281179 N * s0undt3ch_ s0undt3ch 1160282003 Q * Piet_ Ping timeout: 480 seconds 1160282271 J * Piet_ hiddenserv@tor.noreply.org 1160285817 J * dna_ ~naucki@p54BCF83B.dip.t-dialin.net 1160286387 Q * dna_ Quit: Verlassend 1160290743 J * starlein star@fo0bar.de 1160293092 J * soatola ~unlikely@82.153.199.68 1160293336 Q * _are_ Ping timeout: 480 seconds 1160293346 J * Markus23 ~Markus@86.59.55.224 1160294286 J * meandtheshell ~markus@85-124-37-67.dynamic.xdsl-line.inode.at 1160296279 J * bonbons ~bonbons@83.222.36.111 1160296513 Q * Markus23 Remote host closed the connection 1160298690 J * Aiken ~james@tooax8-211.dialup.optusnet.com.au 1160301334 J * Piet__ hiddenserv@tor.noreply.org 1160301743 Q * Piet_ Ping timeout: 480 seconds 1160302642 Q * Piet__ Remote host closed the connection 1160302765 Q * ag- Quit: BRB 1160303629 J * ag- ~ag@caladan.roxor.cx 1160304177 Q * mire Quit: Leaving 1160304847 J * marcfi1 ~mef@c-68-39-177-97.hsd1.nj.comcast.net 1160306402 M * phedny how do I recover from a situation where I cannot start, not stop a vserver and vserver-stat says this: 1160306405 M * phedny 17 1 0 0 0m00s00 0m00s10 14m11s45 1160306589 Q * shedi Quit: Leaving 1160306596 P * marcfi1 1160306645 M * daniel_hozac vkill -c 17 -s 15 1160306666 M * daniel_hozac wait a bit, and then run with 9 instead of 15 if it's still not dead. 1160306738 M * phedny root@aardbei:~# vkill -c 17 -s 15 1160306738 M * phedny vkill: vc_ctx_kill(): No such process 1160306770 M * daniel_hozac hmm. 1160306775 M * daniel_hozac what does ls -l /proc/virtual say? 1160306802 M * phedny dr-xr-xr-x 2 root root 0 2006-10-08 13:27 49153 1160306802 M * phedny -r--r--r-- 1 root root 0 2006-10-08 13:27 info 1160306802 M * phedny -r--r--r-- 1 root root 0 2006-10-08 13:27 status 1160306811 M * phedny 49153 is another vserver i've running ;) 1160306830 M * daniel_hozac so the other guest isn't running anymore, is it? 1160306837 M * daniel_hozac does vps auxc show it? 1160306839 M * phedny vserver-stat does show it 1160306847 M * daniel_hozac vserver-stat is kind of a hack ;) 1160306853 M * phedny root 18002 17 0.0 0.0 0 0 pts/1 Z 13:06 0:00 rc 1160306874 M * phedny along a lot of other lines ;) 1160306878 M * daniel_hozac and xid 17 is this mystery guest? 1160306889 M * phedny yes, it is 1160306940 M * harry phedny: what kernel do you use? 1160306948 M * phedny Linux aardbei 2.6.16.20-vs2.1.1-rc22-ipv6 #1 SMP PREEMPT Thu Jun 22 16:36:09 CEST 2006 i686 GNU/Linux 1160306966 M * daniel_hozac hehe, you might want to update that ;) 1160306966 M * harry now that sounds dutch!? 1160306997 M * phedny daniel_hozac: this is a known problem that has been fixed? 1160306998 M * daniel_hozac i don't recall when we last fixed some reaper issues... 1160307024 M * daniel_hozac hmm, looks like it was around rc19. 1160307064 M * phedny harry: yes, I'm Dutch and my computers at home are named after pieces of fruit :) 1160307074 M * harry ge zou het niet zeggen, he :) 1160307083 M * harry van waar ergens?, toch niet van nederland? ;) 1160307093 M * harry toch wel hoor 1160307098 M * harry HOLLANDER! :p (ikke belg) 1160307103 M * phedny had ik al door :) 1160307113 M * phedny but let's keep it English here :) 1160307122 M * harry anywya, back to trouble shooting in english 1160307124 M * harry indeed 1160307137 A * harry gotta run tough... sry, cant help now 1160307163 M * phedny :) 1160307265 M * daniel_hozac phedny: would be very interesting if you could try to reproduce this on a more recent kernel. 1160307281 M * daniel_hozac you are 15 rcs behind :) 1160307291 M * phedny hehe :) 1160307317 M * phedny hmm, let me try something 1160307381 M * phedny okay, I now figured out how I got into this situation 1160307397 M * phedny because I wanted to stop multiple vservers, I did: vserver mail stop & 1160307410 M * phedny so it stops in the background and I can command other srevers to stop 1160307420 M * phedny but server didn't stop, so I did: vserver mail enter 1160307424 M * phedny and then did: kill -9 -1 1160307436 M * phedny just to try what happened :) 1160307450 M * daniel_hozac can you reproduce it consistently? 1160307467 M * phedny yes, I have ctx 18 in the same state right now 1160307475 M * daniel_hozac interesting... 1160307484 M * daniel_hozac i assume you're using the plain initstyle? 1160307501 M * phedny don't know, I use the default setting after vserver build with debootstrap method 1160307511 M * daniel_hozac hmm, what would be sysv. 1160307515 M * daniel_hozac s/what/that/ 1160307574 M * daniel_hozac just FYI, the stop script will wait 30 seconds before it goes around killing all the processes :) 1160307616 M * daniel_hozac but now that you know how to reproduce it, an update to a more recent kernel would be even more interesting. 1160307638 M * phedny let's try that :) 1160308393 M * phedny this 30 seconds timeout doesn't work here 1160308418 M * phedny all processes I think are related to shutting down are stopped when I start "vserver .. stop" in the background: 1160308422 M * phedny root 5545 2.6 0.8 4364 2128 pts/0 T 13:53 0:00 /bin/bash /usr/sbin/vserver mysql stop 1160308425 M * phedny root 5553 0.0 0.0 96 12 pts/0 T 13:53 0:00 /usr/lib/util-vserver/lockfile /var/lock/vserver.etcvserversmysql.startu 1160308428 M * phedny root 5576 0.0 0.4 4364 1256 pts/0 T 13:53 0:00 /bin/bash /usr/sbin/vserver mysql stop 1160308431 M * phedny root 5577 0.0 0.0 104 24 pts/0 T 13:53 0:00 /usr/sbin/vwait --timeout 30 --status-fd 3 18 1160308466 M * phedny when I issue "fg 1" shutdown finishes cleanly 1160309098 M * daniel_hozac why are the processes showed as stopped? 1160309117 M * daniel_hozac that just shouldn't happen. 1160309270 M * daniel_hozac what happens if you kill -CONT them all? 1160309278 M * phedny linux2.6.18 + vs2.1.1-rc37 is being compiled at the moment, I'll notify when finished and I've tried the same procedure on new kernel 1160309309 N * Bertl_zZ Bertl 1160309314 M * Bertl morning folks! 1160309317 M * daniel_hozac stopping guests in the background works fine for me. 1160309320 M * daniel_hozac morning Bertl! 1160309342 M * phedny daniel_hozac: then shutdown finishes cleanly 1160310429 M * Bertl sidenote: strace also 'stops' processes 1160310462 M * daniel_hozac yeah, but strace causes all sorts of strange behaviour when running the vserver scripts. 1160310484 M * daniel_hozac (meaning: don't do that :)) 1160310676 M * Bertl doener: ping? 1160311090 J * Claw ~Claw@M386P002.adsl.highway.telekom.at 1160311306 M * Bertl welcome Claw! 1160311344 Q * Claw Quit: Life is too short... 1160314256 Q * Aiken Ping timeout: 480 seconds 1160315001 M * phedny new kernel finally finished compiling and has booted :) 1160315010 M * Bertl congrats! 1160315086 M * phedny but problem with "vserver .. stop" is not solved 1160315151 M * phedny daniel_hozac: with newest kernel version, things are exactly the same 1160315171 M * phedny however, by doing 'kill -CONT' on the four processes, shutdown is finished 1160315497 M * daniel_hozac phedny: what shell are you using? and you're just doing vserver ... stop &? 1160315504 M * Bertl may I ask a few questions? 1160315530 M * phedny daniel_hozac: I'm using bash 1160315537 M * phedny Bertl: of course! 1160315554 M * Bertl the guest is running fine, no process inside is stopped, yes? 1160315560 M * phedny indeed 1160315573 M * Bertl okay, you are logged on to the host, via ssh? 1160315578 M * phedny true 1160315594 M * Bertl you have a terminal there (check with tty)? 1160315604 M * phedny root@aardbei:~# tty 1160315607 M * phedny /dev/pts/2 1160315628 M * phedny (btw, I can move to the system and try a xterm) 1160315629 M * Bertl okay, fine, now the line you do is like this: vserver xxx stop ? 1160315630 M * Loki|muh whow, the util-vserver-changelog reads like lots of work being done. 1160315641 M * phedny is like eg: vserver mysql stop & 1160315643 M * Bertl Loki|muh: indeed 1160315653 M * Bertl phedny: ah, you actually background it? 1160315664 M * phedny yes, everything works fine if not in background 1160315673 M * Bertl daniel_hozac: in this case the process loses it's I/O term 1160315681 M * Bertl daniel_hozac: which automatically suspends it 1160315688 M * phedny so it's a userland problem? 1160315704 M * Bertl let's try something like nohup vserver mysql stop & 1160315734 M * Bertl maybe even like this: 1160315738 M * daniel_hozac Bertl: why would it lose its terminal? 1160315739 M * phedny everything fine then :) 1160315746 M * phedny with nohup all problems gone ;) 1160315795 M * Bertl daniel_hozac: that is something you or ensc might answer :) 1160315811 M * Bertl daniel_hozac: it's the way how interactive commands react on backgrounding :) 1160315813 M * daniel_hozac phedny: what util-vserver version? 1160315832 M * phedny vserver 0.30.210 -- manages the state of vservers 1160315840 M * daniel_hozac phedny: 0.30.210 vanilla? 1160315845 M * phedny Ubuntu package 1160315868 M * daniel_hozac Bertl: didn't know that, will have to look in to it. 1160315912 M * daniel_hozac phedny: could you try 0.30.211? i don't think it should fix anything though. 1160315954 M * daniel_hozac hmm, stopping my FC6 guest in the background killed my shell. 1160315965 M * ensc 'setsid vserver ... ' or perhaps '< /dev/null' might help already 1160316042 M * phedny is there a .deb for 0.30.211? as policy for this machine describes I may only install Debian/Ubuntu .deb packages 1160316074 M * phedny with apt-get I only get 0.30.210 1160316103 M * daniel_hozac phedny: not yet, micah uploaded it yesterday though. 1160316157 M * daniel_hozac ensc: btw, will you update the Extras package? 1160316173 M * ensc daniel_hozac: yes 1160316221 M * daniel_hozac ok, FC5 and devel? 1160316248 M * ensc yep 1160316259 M * daniel_hozac nice, thanks. 1160316259 M * ensc but I am not sure about devel and vyum (patches) 1160316270 M * daniel_hozac well, i've used it a little bit. 1160316275 M * daniel_hozac it seems to work. 1160316276 M * ensc I do not have a machine with FC6/devel yet 1160316303 M * daniel_hozac the 2.9.6 patch applies to 3.0 as well. 1160316317 M * phedny when someone can send me .deb for 0.30.211 I'll be happy to help test it 1160316327 M * phedny then for now I'll continue installing webserver vserver for work ;) 1160316406 M * daniel_hozac phedny: http://incoming.debian.org/ 1160316461 M * phedny hmm, didn't know that site :) 1160316474 M * daniel_hozac ensc: oh, i just realized the 3.0 will mess up the version test. 1160316505 M * daniel_hozac phedny: neither did i, google is my friend :) 1160316516 M * phedny hehe :) 1160316968 M * phedny in 0.30.211 problem isn't solved, but something different happens 1160317125 M * daniel_hozac different? 1160317166 M * phedny no, nothing different 1160317170 M * phedny I did something different :) 1160317201 M * daniel_hozac hehe. 1160317203 M * phedny still, sending CONT signal to the four processes makes the shutdown process finish 1160317209 M * daniel_hozac so do ensc's tips make it work? 1160317221 J * shedi ~siggi@inferno.lhi.is 1160317226 M * phedny with setsid? 1160317246 M * daniel_hozac and/or the < /dev/null 1160317333 M * shedi am I supposed to be able to assign the same loopback interface to all my guest servers on a single host 1160317335 M * phedny with setsid it works, with then that should do what you want, no (daniel_hozac ?) 1160322931 M * kichukov right Bertl ;-) 1160322934 M * kichukov will try ;-p) 1160322936 M * daniel_hozac no, but it will tell you to set nodev. 1160322950 M * daniel_hozac (i suppose that could be improved) 1160322953 M * Bertl daniel_hozac: ah, so that is a bug then? 1160322980 M * Bertl I always thought that --interface implied nodev unless a device is specified 1160323017 M * daniel_hozac IIRC not specifying dev will cause an error. 1160323036 M * Bertl hmm, that must be new then ... 1160323065 M * Bertl (or my memory is wrong) 1160323094 M * daniel_hozac test -n "$dev" -o -e "$iface"/nodev || { 1160323094 M * daniel_hozac echo $"No device specified for '$iface'" >&2 1160323095 M * daniel_hozac return 1; 1160323124 M * daniel_hozac so you should get that when you try to start the guest. 1160323300 M * Loki|muh uh thats bad. 1160323323 M * Loki|muh so everyone has to change the configs? 1160323366 M * Loki|muh ups, forget what I said 1160323371 M * Loki|muh nvm 1160323460 M * daniel_hozac this hasn't changed since january 2005. 1160323521 M * Loki|muh yeah, I'm too blind for my own config :( 1160323570 J * mauro ~mauro@host164-178-dynamic.5-87-r.retail.telecomitalia.it 1160323808 M * Bertl welcome mauro! 1160323822 M * mauro Hi Bertl! 1160323840 M * azazel toh, un'italiano:) 1160323852 M * mauro azazel: :) 1160323922 J * malveo ~malveo@host164-178-dynamic.5-87-r.retail.telecomitalia.it 1160323951 M * malveo due it!!! 1160323979 M * Bertl malveo, mauro, azazel: channel language is english 1160324005 M * malveo ok Bertl, thx 1160324010 M * Bertl np 1160324021 M * azazel Bertl: yes, we know:) 1160324154 Q * malveo 1160324725 Q * mauro Quit: bye bye! 1160324821 M * kichukov hmmm, guys 1160324848 M * kichukov so having --interface and no --netdev will yield error upon guest startup ? 1160324856 M * daniel_hozac right. 1160324894 J * marcfi1 ~mef@c-68-39-177-97.hsd1.nj.comcast.net 1160324899 M * marcfi1 hello 1160324924 M * kichukov so daniel_hozac how can i work around that issue then ? 1160324927 M * kichukov hello marcfi1 1160324944 M * daniel_hozac touch the nodev file after you've built it? 1160324951 N * marcfi1 marcfiu 1160324957 M * marcfiu hello Kichukov 1160324971 M * kichukov daniel_hozac, will do 1160324971 M * kichukov ;-) 1160324972 M * kichukov tnx 1160325010 M * kichukov figured out i need no dev anymore, since i had to remove one of the nics and the whole system hang upon startup(host startup) ? 1160325046 M * kichukov i had the ip bound to that device 1160325121 M * doener Bertl: pong! 1160325132 M * doener lost connectivity yesterday... back at home now 1160325344 M * Bertl doener: np 1160325407 P * marcfiu 1160325511 M * kichukov daniel_hozac, clever enough: vserver vn build -m skeleton --initstyle sysv --hostname vn.xx --context 111 --interface 192.168.10.1/32 1160325511 M * kichukov No device specified for interface '0'; do not forget to set the 'nodev' option 1160325516 M * kichukov it warns you;-) 1160325586 M * daniel_hozac like i said ;) 1160325607 M * kichukov aye 1160325608 M * kichukov cool 1160325662 M * kichukov now that i have the skeleton, can i cp old_guest/ new_guest/ && vserver new_guest start 1160325694 M * daniel_hozac yep. 1160325733 M * kichukov cool 1160325741 M * kichukov thanks for the assistance 1160325990 M * daniel_hozac is the mailing list down? the mail i sent like 6 hours ago still hasn't gotten back to me. 1160325995 M * daniel_hozac nor to the archives. 1160326115 M * Bertl I received this one: Date: Sun, 8 Oct 2006 15:08:10 +0200 1160326145 M * Bertl (via the mailing list, so I'd assume it is still working) 1160326180 M * daniel_hozac hmm, ok. i haven't gotten that one yet. 1160326321 M * daniel_hozac "We have received some recent bounces from your address. Your current bounce score is 3.0 out of a maximum of 5.0." :/ did it get put on some blacklist again? 1160326437 J * _node node@c-69-143-148-254.hsd1.md.comcast.net 1160326484 M * kichukov daniel_hozac, whose address? 1160326501 M * daniel_hozac my address. 1160326510 M * kichukov alright ;-/ 1160326668 M * kichukov okay, now that i started the guest, i am getting that error: 1160326672 M * kichukov Starting apache 2.0 web server...apache2: Could not determine the server's fully qualified domain name, using 127.0.0.1 for ServerName 1160326672 M * kichukov (99)Cannot assign requested address: make_sock: could not bind to address 192.168.10.1:9080 1160326672 M * kichukov no listening sockets available, shutting down 1160326689 M * kichukov basicly i have no interface for the guest 1160326719 M * daniel_hozac so ip a inside the guest doesn't list anything? 1160326762 M * kichukov indeed 1160326776 M * daniel_hozac well then, that won't work :) 1160327071 M * kichukov aye 1160327081 M * kichukov soatola, a workaround could be ? 1160327305 M * Bertl is the address listed on the host? 1160327313 M * doener kichukov: did you configure that address on the host 1160327320 M * doener Bertl: 2fast4me 1160327343 M * doener Bertl: feeling better by now? or still ill? 1160327661 M * kichukov i did not 1160327663 M * kichukov should i ? 1160327672 M * doener yeah, that's what "nodev" means 1160327681 M * kichukov okay 1160327693 M * kichukov how to assign that address to the host and specify no interface? 1160327695 M * doener nodev stops the tools from adding and removing the address on vserver startup 1160327707 M * kichukov maybe that is a FAQ but still ... 1160327724 J * dreamind ~dreamind@C2107.campino.wh.tu-darmstadt.de 1160327735 M * doener ... and stop (removing on startup would be pretty "interesting" ;) 1160327761 M * dreamind Hi :) 1160327766 M * doener welcome dreamind 1160327782 M * dreamind Hi doener :) 1160327808 J * michael ENETDOWN@fw-ext.konaktiva.tu-darmstadt.de 1160327825 M * michael hi 1160327830 M * doener welcome michael 1160327869 M * kichukov ip addr add 192.168.10.1 1160327869 M * kichukov Not enough information: "dev" argument is required. 1160327880 M * daniel_hozac yep. 1160327880 M * dreamind Hi michael :) 1160327883 M * doener ... dev eth0 (for exmaple) 1160327891 M * doener s/ma/am/ 1160327894 M * kichukov well yeah 1160327896 M * kichukov okay 1160327919 M * kichukov i got that, it still has to be assigned to a device 1160327926 M * daniel_hozac of course. 1160327929 M * kichukov the address cannot be floating around ;-) 1160327932 M * kichukov alright 1160327948 M * dreamind well I've just asked in #debian.de, but maybe somebody here knows an answer... I want to have local changes to the /etc/services file, so does anybody know a way I can have the system look also in /etc/services.local (for example)? 1160327951 M * kichukov what i initially thought was too much virtualized;-) 1160327965 M * daniel_hozac dreamind: why not change /etc/services? 1160327979 M * dreamind daniel_hozac: because on every debian upgrade, I have to manually fix it. 1160327988 M * daniel_hozac dreamind: that sounds like a Debian bug to me. 1160327997 M * dreamind I don't see it as a bug. 1160328007 M * dreamind its a config file, and you are asked, what you want to do. 1160328009 M * daniel_hozac overwriting configuration files on updates? 1160328017 M * dreamind its _not_ automatically overwritten. 1160328022 M * dreamind I wroute you have to manually fix it. 1160328025 M * daniel_hozac well then, what's the problem? 1160328035 M * dreamind well just forget it. 1160328040 M * kichukov night all, going home nnow;-) 1160328042 Q * kichukov Quit: Leaving 1160328062 J * mire_ ~mire@76-167-222-85.adsl.verat.net 1160329246 Q * fosco Remote host closed the connection 1160329334 J * smartman_ ~sdfsdf@85.187.125.155 1160329340 M * smartman_ anybody with experience in online education and online degrees ? 1160329360 M * Bertl smartman_: wrong channel :) 1160329394 M * smartman_ i am looking for online linux education 1160329506 M * harry smartman_: start by disassembling 3 compilers of choice 1160329514 M * harry then: rewrite the linux kernel 1160329518 M * harry and you're off :) 1160329523 M * harry (or: just use it ;)) 1160329546 M * harry so sry... can't help you... except: use it :) 1160329574 M * smartman_ :) 1160329575 M * smartman_ ok 1160329767 Q * lilalinux Remote host closed the connection 1160331056 J * _are_ are@dslb-084-057-205-123.pools.arcor-ip.net 1160331143 M * _are_ hi 1160331152 M * Bertl hey _are_! 1160331670 M * dreamind hm, is there a way to have vserver ... enter change to /root ? 1160331800 M * _are_ just type cd (ok, probably not the solution you asked for ;) 1160331810 M * dreamind _are_: not really ;) 1160331821 M * Bertl dreamind: why /root ? 1160331831 M * dreamind Bertl: well, I'm used to it ;) 1160331851 M * Bertl ah, what about putting that into some bashrc/profile? 1160331859 M * dreamind uh, thats a hack. 1160331883 M * Bertl hmm, no it would be a hack if vserver .. enter did that :) 1160331925 M * _are_ dreamind: vserver ... enter gets you into the verser, what you want is 'log into the vserver' 1160331937 M * dreamind _are_: yup 1160331944 M * Bertl use ssh for that 1160331968 M * _are_ as you have no vserver-internal utmp entry wuith vserver entry and limited access to the pty you come from, this is not the best idea anyway -> use ssh 1160331973 M * dreamind hm, well I don't want to have ssh in the vserver, my root login in the vserver doesn't have a passwort (and it has login disabled) 1160331991 M * daniel_hozac _are_: 0.30.211 allocates a new pty for vserver ... enter. 1160331992 M * dreamind well I did put it in my zshrc 1160332013 M * _are_ daniel_hozac: nice, i guess same for ... exec? 1160332022 M * daniel_hozac no, not for ... exec. 1160332049 M * _are_ to bad, I'd loved to be able to call aptitude that way 1160333031 Q * smartman_ 1160333756 J * lukasgraf ~lukas@80-218-116-45.dclient.hispeed.ch 1160333852 M * sp evening 1160333881 M * sp I just wanted to drop by and ask what the Vserver_Hosting wiki page is about 1160333916 M * sp it looks as if only companies using linux-vserver are listed there, but I'm part of a, let's call it, 'bigger' project using linux-vserver as well 1160333941 M * Bertl we have a second page, called vserver users 1160333960 M * Bertl it wasn't migrated to the new wiki yet 1160333994 M * sp ah, okay 1160333996 M * Bertl basically it allows for groups and individual who _use_ Linux-VServer but do not provide any hosting services 1160334025 M * Bertl as the old wiki is in lockdown mode 1160334028 M * sp going to watch it for now then add myself (am using Linux-VServer for myself as well) and the project then 1160334052 M * Bertl best would be to send me an email, I'll move that page soon, and I can add you then 1160334180 M * sp sounds like a plan 1160334201 M * Bertl okay, you know how to reach me? 1160334210 M * sp doubt it's hard to find out 1160334226 M * sp done 1160334231 M * sp I'll drop you a mail now 1160334235 M * Bertl great, tx! 1160334350 M * lukasgraf Hello everybody! 1160334358 M * lukasgraf I'm using vserver on Gentoo, Kernel 2.6.15-vs-2.0.1-r5. When I try to copy an existing vserver to a new one using "vcopy", I get these errors: 1160334373 M * lukasgraf rsync: writefd_unbuffered failed to write 4 bytes: phase "unknown": Broken pipe 1160334379 M * lukasgraf rsync error: error in rsync protocol data stream (code 12) at io.c(666) 1160334392 M * lukasgraf First I though this might be caused by the vserver still running, but stopping it before attemping to clone it doesn't make a difference. 1160334417 M * lukasgraf I also checked the filesystem (ext3), e2fsck says it's clean. So, what else could be a cause for these errors? 1160334447 M * Bertl hmm, at first glance I'd say rsync is failing somehow 1160334480 M * Bertl could you verify that rsync is working properly by copying a test directory on the host? 1160334485 M * lukasgraf Yep, that's also my impression. But the error messages are somewhat disturbing, and unfortunately don't reference any specific files.. 1160334497 M * lukasgraf Yes, I'll do that.. 1160334531 M * Bertl btw, you should update to a more recent kernel soon# 1160334543 M * lukasgraf I know :-) 1160334596 M * lukasgraf Have been putting that up for too long, because "it works"... I'm still a little bit scared of kernel updates and reboots on my remote root server ;-) 1160334710 M * lukasgraf Ok, rsync seems to run just fine... 1160334739 M * lukasgraf Also, because I'm running Gentoo, which heavily depends on rsync, I'd probably notice a "general" rsync failure very early 1160334755 M * Bertl okay, as I don't know vcopy, check if it has some '--debug' option? 1160334792 M * lukasgraf Mhh, not that I know of, it doesn't even have a usage... 1160334817 M * Bertl hmm, then it's probably time to file a bug report to gentoo 1160334821 M * lukasgraf (util-vserver 0.30.210-r13) 1160334860 M * Bertl but in the meanwhile, you can use that rsync to copy it manually 1160334881 M * Bertl i.e. first create a skeleton guest with the necessary options 1160334884 M * lukasgraf just /vserver/ and /etc/vserver/, right? 1160334892 M * lukasgraf oh, ok 1160334907 M * Bertl then rsync -axzH --numeric-ids --progress /guest/A /guest/B 1160334923 M * Bertl you can add --dry-run to check if that would do what you want 1160334964 M * lukasgraf I guess, I'll encounter the same rsync error there, but hopefully it will be more verbose / specific 1160334971 M * lukasgraf Thanks for the advice! 1160335001 M * Bertl you're welcome! but if rsync gives you an error there, then rsync needs to be fixed :) 1160335016 M * lukasgraf Hmm... 1160335026 M * lukasgraf It complains about a "Broken Pipe".. 1160335071 M * lukasgraf I've had a similar error with CVS post-commit hooks, which were caused if STDOUT of a script wasn't read *completely* by the script it was piped to.. 1160335072 M * sp Bertl: what I've been wondering about is if there's a 'clean' way of using virtual interfaces for the vservers 1160335126 M * Bertl sp: what kind of virtual interfaces do you have in mind? 1160335129 M * sp as in having some sort of 'host-only' networking via a virtual switch running in userspace 1160335153 M * sp well, it doesn't have to run in userspace, that's just the way I've used uml before 1160335163 M * Bertl sp: ah, well, we avoid that (and network virtualization) because of the intrinsic overhead 1160335188 M * Bertl sp: Linux-VServer does network (ip) isolation instead of virtualization 1160335229 M * sp I know how it works right now, and I'm well-aware of the overhead this would generate 1160335266 M * sp it's just a feature I'd need on one of my systems and I'm wondering about if it would be worth hacking in (or at least trying to do so) 1160335303 M * Bertl well, mainline (kernel.org) is working on network virtualization too, and I can assure you, it's not that easy 1160335332 M * sp didn't want to say it's easy either 1160335341 M * Bertl if you need a virtual stack and don't mind the overhead, the best way is probably Xen 1160335400 M * sp to be honest, I got used to linux-vserver and am thinking about using tap devices for something like that 1160335462 M * Bertl well, I doubt you can hack that into the networking stack, but please prove me wrong there ... 1160335630 M * sp I'm just thinking about it right now, but in general I should be able to access the tap device inside the 'guest' in a similar way openvpn for example does 1160335662 M * Bertl yes, but the problem is, the tun/tap devices connect _the_ network stack with userspace, right? 1160335679 J * BeLu982 B.Lukas@mail.openvcp.org 1160335680 N * BeLu982 BeLu 1160335691 M * sp mhh, good point 1160335695 M * Bertl now let's assume you have two tap/tun devices connected together in userspace 1160335721 M * Bertl how do you tell the network stack to send a packet destinated to tun2 into tun1 in _the_ network stack? 1160335757 M * Bertl don't forget, UML is in userspace and has an userspace network stack 1160335810 M * sp good point, again 1160336328 J * root_ ~root@tor-irc.dnsbl.oftc.net 1160336354 J * fosco fosco@konoha.devnullteam.org 1160336365 M * root_ Muslims are terrorists 1160336373 M * root_ and their women are not submissive 1160336378 M * root_ we should kill their women 1160336389 Q * BeLu Ping timeout: 480 seconds 1160336403 M * root_ how dare they wear a veil (which means submission in Christianity) and be bitches 1160336411 M * root_ I say no to women's rights 1160336425 M * root_ and I say KILL to muslim women and all feminist women 1160336433 M * root_ because they want to have rights too 1160336436 M * root_ anyone agree 1160336453 M * Bertl root_: wrong channel 1160336493 M * root_ Bertl, what channel should I go to, #debian doesn't allow tor? 1160336505 M * Bertl try #offtopic 1160336518 M * Loki|muh we had such a guy here the day before yesterday 1160336526 M * root_ Loki|muh, really? 1160336528 M * root_ it wasn't me 1160336531 M * root_ what did he say? 1160336550 M * Loki|muh he flames something about #debian not wanting him :p 1160336564 M * root_ did he say anything about women's rights 1160336566 M * root_ or muslims 1160336572 M * Loki|muh no 1160336576 M * root_ ah 1160336576 M * Bertl root_: so what do you need/want to discuss regarding Linux-VServer? 1160336602 M * root_ oh, it doesn't have good performance on a patched linux kern with grsecurity 1160336623 M * root_ I was wondering if maybe you could work with them to up the preformance? 1160336632 M * Bertl really, guess you need to talk to harry then 1160336648 M * Bertl on vanilla kernels the performance is very good :) 1160336669 M * root_ well there is a new thing in the grsec patch which sometimes affects VMs etc 1160336693 M * root_ sometimes it affects them, sometimes it doesn't 1160336702 M * root_ I'm not too sure why though 1160336703 M * Bertl and what's that? 1160336724 M * root_ it's one of their new security features that does something with the memory 1160336732 M * root_ the 2 new features are: 1160336742 M * root_ whenever a page is outdated it flushes it out 1160336762 M * root_ (so that info can't be discovered from ram) 1160336774 M * root_ and ... I think some randomization thing 1160336879 M * root_ you can get grsec at grsecurity.org 1160336882 M * root_ and test with it 1160336896 M * Bertl well, shouldn't that affect vanilla + grsec too? 1160336934 M * root_ yes 1160336942 M * root_ grsec is a kernel patch 1160336956 M * Bertl okay, so it is a feature/issue for grsec then 1160336965 M * Bertl and not really Linux-VServer related 1160336998 M * root_ well grsec is a security patch (that is used alot) and if the 2 projects could make things work happily with the security... would be nice 1160337057 M * root_ the description of the option suggested that the performance hit could be avoided by the vm 1160338149 M * dreamind hm, is there any easy way to allow a vserver to bind to any ip of a specific subnet? 1160338168 M * Bertl not yet, but there will soon be 1160338208 M * dreamind like I'm currently using a setup with several apache (and some other webservers) running at some ips of the 127.0.0.0/8 subnet, and a reverse proxy which forwards the requests accordingly 1160339148 Q * meandtheshell Quit: exit (0); 1160340492 M * Bertl okay, I'm off for today ... maybe back later ... 1160340500 M * Bertl have a good one everyone ... cya! 1160340510 N * Bertl Bertl_zZ 1160340670 Q * bonbons Quit: Leaving 1160341138 J * besonen ~besonen@dsl-db.pacinfo.com 1160341368 J * Aiken ~james@tooax6-209.dialup.optusnet.com.au 1160342649 N * hardwire-afk hardwire 1160344216 Q * soatola Quit: soatola 1160345963 J * transaci1 ~transacid@transacid.de 1160346037 Q * transacid Ping timeout: 480 seconds 1160347744 M * hardwire blah 1160347759 M * doener sounds reasonable ;) 1160347785 M * hardwire painting today 1160347823 M * hardwire so.. does the vserver patch to the kernel know 100% what memory allocation the vserver context has? 1160347861 Q * dna_ Quit: Verlassend 1160347869 M * daniel_hozac no. 1160347874 M * hardwire :( 1160347880 M * daniel_hozac it knows how much though. 1160347897 M * hardwire I can't quite understand how its chroot++ basically 1160347897 M * doener daniel_hozac: what about shared pages? 1160347900 M * hardwire but I am getting there 1160347935 M * doener hardwire: easy :) chroot breakouts need access to a dir "above" the current chroot 1160347937 M * daniel_hozac doener: hmm, how do we account thsoe? 1160347946 M * daniel_hozac s/thsoe/those/ 1160347961 M * daniel_hozac doener: do we just account them to the first context that use them? not at all? 1160347972 M * doener hardwire: if you chroot ("/foo"); cd ("../../../..") you escape the chroot 1160347984 M * hardwire so thats primarily why there is no suspend-state mechanism 1160348022 M * doener hardwire: the vserver patch adds a special flag that can be set on a directory and completely stops you from touching or passing that directory 1160348029 M * hardwire barrier 1160348046 M * doener in chdir ("../../../..") one ".." is eg. /vservers 1160348076 M * doener while the kernel resolves the dotdots, it checks for each directory if you may access it, and if not the whole command fails 1160348119 M * hardwire hah 1160348122 M * hardwire well 1160348127 M * hardwire I never tried that before 1160348137 M * doener as you _must_ pass the directory with the barrier set to get upwards, there's no way out 1160348137 M * hardwire cp ../../../../etc/fstab ./tmp/ from chroot only 1160348139 M * hardwire kinda creepy 1160348140 M * hardwire but 1160348145 M * hardwire kinda helpfull for other things I am doing' 1160348165 M * hardwire so the only thing that cares about barrier is a running vserver context 1160348175 M * hardwire since chroot still doesn't care 1160348227 M * doener no, the context is a passive thing ;) being in a context just triggers the barrier check 1160348258 M * doener and if it would always be active, you couldn't reach the vservers at all, would be quite hard to start them then ;) 1160348286 M * doener and for example secure-mount relies on breaking out of a chroot 1160348298 M * doener daniel_hozac: no idea 1160348340 M * daniel_hozac hmm, secure-mount relies on breaking out of a chroot? really? 1160348384 M * doener daniel_hozac: check restoreRoot 1160348415 Q * lukasgraf Quit: Leaving 1160348463 M * daniel_hozac hmm, ok. 1160348498 M * hardwire secure-mount? 1160348890 M * doener part of util-vserver that handles mounts in vservers (fstab) 1160348917 M * doener it chroots, resolves paths etc. to avoid symlink attacks 1160348959 M * hardwire woo 1160348961 M * hardwire I mean it 1160348964 M * hardwire Woo! 1160348967 M * hardwire ok 1160348971 M * hardwire I have to go chase a kitty 1160349450 Q * dreamind Quit: dreamind 1160350127 Q * root_ Remote host closed the connection 1160351237 J * root_ ~root@tor-irc.dnsbl.oftc.net