1300762779 J * bsingh ~balbir@122.172.25.100 1300768576 J * bwh ~ben@shadbolt.e.decadent.org.uk 1300768625 M * bwh Just looking at 1300768646 M * bwh It's kind of my fault, as I've changed an upstream fix to avoid an ABI change 1300768698 M * bwh However, the more I look at it, I think the upstream fix for CVE-2010-4243 needs to be extended for VServer 1300768824 M * bwh Upstream fix is commit 3c77f84, and it adjusts mm->total_vm which I understand should not be done directly in VServer 1300769062 Q * FireEgl Ping timeout: 480 seconds 1300769473 Q * dannf Ping timeout: 480 seconds 1300771302 M * Bertl bwh: so, does it happen on a mainline kernel as well? 1300771414 M * Bertl and yes, if that fix changes vm handling, you need to do that in a way Linux-VServer knows about the accounting changes, otherwise you'll end up with the described issue 1300771617 M * bwh Well the mainline fix avoids it because it adjusts total_vm up, then down later 1300771637 M * bwh I screwed it up by deleting the second adjustment 1300771665 M * Bertl ah, I see ... 1300771680 M * bwh I don't think the second adjustment is needed because we're about to free the mm 1300771704 M * Bertl well, both should be made in a Linux-VServer aware way, if you want to use the 'old' vm accounting 1300771712 M * bwh *but* then VServer will subtract total_vm from the server's quota, when the original addition wasn't accounted 1300771729 J * FireEgl FireEgl@2001:470:e056:1:e807:88b:acdd:f268 1300771772 M * bwh Bertl: So I think you should be patching acct_arg_pages() 1300771813 M * Bertl probably on the next patch update, I'll do ... but recent kernels do not use that accounting anymore 1300771891 M * bwh What do you mean? 1300771906 M * bwh 2.6.38 has acct_arg_size() 1300771924 M * Bertl but 2.6.38 doesn't have Linux-VServer vm accounting anymore :) 1300771948 M * bwh Well no mainline kernel has it... 1300771970 M * Bertl no Linux-VServer patch for 2.6.38 has it either 1300771986 M * bwh Are you using cgroups memory controller instead? 1300771993 M * Bertl correct 1300772017 M * bwh But is the patch for 2.6.32 maintained? 1300772042 M * Bertl partially 1300772077 M * Bertl i.e. it's kind of on-demand ... 1300773388 J * arekm arekm@carme.pld-linux.org 1300773448 Q * bsingh Ping timeout: 480 seconds 1300774175 M * Bertl bwh: but I do not see those changes in 2.6.32.33 for example? 1300774296 Q * _nono_ Read error: Operation timed out 1300774396 J * petzsch ~markus@p57B67A96.dip.t-dialin.net 1300774422 M * bwh Bertl: Oh, true, I need to go through and look for fixes in Debian that haven't gone upstream yet. 1300774443 M * bwh Bertl: maks has covered most of them 1300774477 M * bwh Well, I created my first vserver, reproduced the bug and verified the obvious fix 1300774524 M * bwh http://svn.debian.org/wsvn/kernel/dists/squeeze/linux-2.6/debian/patches/features/all/vserver/vserver-complete-fix-for-CVE-2010-4243.patch 1300774604 M * bwh Bertl: s/upstream/to stable/ 1300774680 M * Bertl I'd suggest to leave the mm->total_vm line there commented out, just to keep in style with the other vm changes 1300774716 M * Bertl (but that's just to simplify future changes for you :) 1300775102 J * _nono_ ~gomes@licencieux.ircam.fr 1300777282 J * ncopa ~ncopa@3.203.202.84.customer.cdi.no 1300777385 M * Bertl off to bed now ... have a good one everyone! 1300777389 N * Bertl Bertl_zZ 1300778983 Q * FireEgl Remote host closed the connection 1300779485 Q * derjohn_mob Ping timeout: 480 seconds 1300780017 J * FireEgl ~FireEgl@173-25-19-139.client.mchsi.com 1300780635 J * derjohn_mob ~aj@213.238.45.2 1300782068 J * bsingh ~balbir@122.172.47.199 1300782963 J * BenG ~bengreen@cpc12-aztw24-2-0-cust146.aztw.cable.virginmedia.com 1300783363 Q * bsingh Ping timeout: 480 seconds 1300783782 M * arekm hmm, I wonder why this doesn't work - http://pastebin.com/vi03LLBt (guest was restarted... actually even manual echo 512M > /dev/cgroup/carme-pld-srv/memory.limit_in_bytes doesn't change what "free" shows 1300783970 M * Mr_Smoke arekm: silly questino but I have to ask ... using VIRT_MEM at all ? 1300783986 M * arekm Mr_Smoke: not that silly, seems not 1300783990 M * Mr_Smoke :) 1300784011 M * arekm works, thanks 1300784015 M * Mr_Smoke Also, even though /etc/vservers//cgroup/* seem to accept stuff like 512M, I'm not sure /dev/cgroup does 1300784018 M * Mr_Smoke Oh it does then 1300784105 J * bsingh ~balbir@122.167.171.61 1300784412 M * BenG Mr_Smoke, yeah it does, I thought I was having trouble with that but it seems to work fine 1300784451 M * BenG anyone know a quick way I can get from having a guests nodename to having it's actually name 1300784458 M * BenG ahem, actual name 1300784470 M * BenG that is the one it gets in vserver-stat 1300784507 M * Mr_Smoke I don't understand your question 1300784543 M * arekm does anyone know how cpu is divided among cgroups limited vs not limited guests? Example, 3 guests, 2 of them have shares==100, 1 of the has no shares set at all. Each guest gets cpu hungry process eating all cores. How cpu will be divided among guests then? 1300784565 M * BenG default cpu.shares is 1024 1300784592 M * arekm oh! 1300784612 M * Mr_Smoke What about the host ? 1300784615 M * BenG Mr_Smoke, I can get the nodename like this for example: cat /proc/virtual//nsproxy | grep Nodename 1300784617 M * arekm also 1024 1300784638 M * Mr_Smoke arekm: really ? 1300784641 M * Mr_Smoke Good to know 1300784669 M * BenG but host processes are at the root cgroup so it's different for scheduling 1300784693 M * BenG so at /dev/cgroup/cpu.shares 1300784706 M * Mr_Smoke BenG: what's with the nodename then ? 1300784728 M * BenG I have some guests were the guest name is different from the nodename 1300784739 M * BenG I have some guests where the guest name is different from the nodename 1300784748 M * BenG and I want to find one from the other 1300784761 M * Mr_Smoke Ok, no idea then 1300784777 M * arekm different but what it means? How 1024 of host relate to guests? 1300784797 J * kir ~kir@swsoft-msk-nat.sw.ru 1300784815 M * BenG arekm, either read the docs or test it, I can't really recall 1300784863 M * BenG but as to the guest themselves, they share in proportion to their cpu.shares 1300784894 M * BenG so if you have 2 guests with 1024 and 512 shares, the first will get 2/3 processing and the other 1/3 1300784923 M * arekm that I know :) googling for host vs guests 1300784987 M * BenG please let me know if you find something conscise and simple, not many people seem to be testing cgroups that much, or at least giving much feedback 1300785007 M * BenG in the mean time, I'm going to try it out 1300785825 M * arekm BenG: my guess is that it will be hierarchical. so guests share whole 1024 of parent. If there were 3-level three like host-2x guest--2x subguest then subguest whould share guest "shares" while guest would share host "shares" etc. Still looking for some confirmation of that 1300785841 M * arekm s/3 level three/3 level tree/ 1300785863 M * BenG yeah I get you, it'll be something like that, I want to try it on the ground though 1300786415 M * arekm but this still doesn't explain what happens to processes that are running in host vs guests 1300786673 M * BenG I'm setting up a test environment, I'll be back with results in about 10 mins 1300786993 M * daniel_hozac arekm: you're right. 1300787075 M * daniel_hozac one host process (unless configured differently) will be given the equivalent share to an entire guest. 1300787119 M * daniel_hozac BenG: grep /etc/vservers/*/uts/nodename? 1300787180 M * arekm daniel_hozac: so for example host shares 1024 will still be equal to guest even if guest has shares set to 2000 ? 1300787222 M * arekm (aka the same just different "unit") 1300787249 M * BenG daniel_hozac, thankyou, I really should have thought of that myself 1300787332 J * ghislain1 ~AQUEOS@adsl2.aqueos.com 1300787348 M * daniel_hozac arekm: if the guest has 2000 shares, it will be equivalent to two host processes. 1300787362 M * daniel_hozac or well, no. 1300787372 M * daniel_hozac that's for the other way around :) 1300787405 M * daniel_hozac well, hmm. 1300787418 M * daniel_hozac i'm confusing myself now. 1300787480 J * MooingLemur ~troy@ipv4.pinchaser.com 1300787516 Q * ghislain synthon.oftc.net osmotic.oftc.net 1300787516 Q * MooingLe1ur synthon.oftc.net osmotic.oftc.net 1300787830 N * ensc Guest1726 1300787839 J * ensc ~irc-ensc@p5DF2EBCE.dip.t-dialin.net 1300787872 M * BenG arekm, what daniel_hozac said works in practice 1300787886 M * BenG the host appears to be just another share 1300787907 M * BenG guest/guest/host will share 1:1:1 by default 1300787928 M * BenG changing guest values redistributes the cycles as you'd expect 1300787972 M * arekm so host 1000, guest 2000 means host gets 1/3 of cpu power in practice? 1300787995 M * arekm regardless of cgroup hierarchy level 1300788000 Q * Guest1726 Ping timeout: 480 seconds 1300788066 M * BenG can you set the guest level? 1300788071 M * BenG I don't seem to be allowed 1300788080 M * BenG sorry, can you set the HOST level? 1300788083 M * BenG I don't seem to be allowed 1300788089 M * BenG I can set guests fine 1300788124 M * arekm same here. So host 1024 (not touched) and guest 2048 1300788195 M * BenG yep, that will work 1300788227 M * BenG I don't seem to be able to create sub directories inside /dev/cgroup/ either 1300788261 M * daniel_hozac what cgroups do you have mounted? not all of them support hierarchies. 1300788269 M * daniel_hozac check /proc/cgroups 1300788335 M * BenG ah right, pretty much all of them 1300788425 A * arekm has to refresh memory (via google) on how to make guest have access to network without having separate ip address 1300788453 M * daniel_hozac give it the host's? 1300788494 M * arekm long time ago it afaik worked without adding any ip somehow, I assume this way no longer exists? 1300788513 M * daniel_hozac depends on what you're referring to. 1300788519 M * daniel_hozac you can give it 0.0.0.0 to give it limitless power. 1300788549 M * daniel_hozac or touch /etc/vservers//noncontext 1300790154 Q * petzsch Quit: Leaving. 1300790898 Q * bsingh Ping timeout: 480 seconds 1300791427 J * ghislain ~AQUEOS@adsl2.aqueos.com 1300791435 J * bsingh ~balbir@122.172.35.136 1300791713 Q * ghislain1 Ping timeout: 480 seconds 1300791910 M * arekm BenG: so my testing also shows that there is no hierarchy for cpu.shares. Each cgroup (regardless of host, guest) gets its cpu.shares/sum-of-all-shares-in-all-cgroups 1300791941 M * BenG okay 1300792017 M * BenG so, it would be useful to have a seperate /dev/cgroup// so that we can still use the hierarchy in other cgroup file systems 1300792181 P * kir Leaving. 1300792744 M * daniel_hozac arekm: that sounds like a bug. do you have group scheduling enabled? 1300792968 Q * bsingh Ping timeout: 480 seconds 1300793125 M * arekm daniel_hozac: yes, http://pastebin.com/vcrFNrgG 1300793168 M * arekm (so now I have no idea what behaviour is expected) 1300793376 J * coredumb ~AndChat@213.55.131.81 1300793507 J * bsingh ~balbir@122.167.174.61 1300793652 Q * coredumb 1300793681 M * daniel_hozac arekm: they're supposed to be hierarchial in that case. 1300793688 M * daniel_hozac or that was my understanding at least. 1300793885 M * BenG so what behaviour would one expect in a hierarchical setup? 1300793913 M * BenG how could I test for it? 1300794421 M * daniel_hozac a host process should be scheduled as often as any guest process. 1300794427 M * daniel_hozac assuming equivalent weights. 1300794578 Q * FireEgl Remote host closed the connection 1300794689 M * daniel_hozac i.e. in your case above, you should have gotten 2:1:1 1300794819 J * petzsch ~markus@p57B67A96.dip.t-dialin.net 1300794997 N * ensc Guest1738 1300795007 J * ensc ~irc-ensc@p5DF2C4C0.dip.t-dialin.net 1300795404 Q * Guest1738 Ping timeout: 480 seconds 1300796880 Q * BenG Quit: I Leave 1300796964 J * FireEgl FireEgl@2001:470:e056:1:18f8:305e:4c4:bd3c 1300797121 M * arekm daniel_hozac: host has 1024, I give guest 2048 and guest gets 66% as shown by vtop (I run cpuburn in host and guest) 1300797157 M * daniel_hozac right, that's as it should be. 1300797279 N * ensc Guest1743 1300797289 J * ensc ~irc-ensc@p5DF2C4C0.dip.t-dialin.net 1300797357 M * petzsch has someone had a chance to test blk_cgroup? sounds like a useful thing to have in a vserver envoirment. 1300797444 M * petzsch link to the documentation of the module: http://bit.ly/fTZZC9 1300797450 Q * Guest1743 Ping timeout: 480 seconds 1300798938 Q * bsingh Ping timeout: 480 seconds 1300799491 J * bsingh ~balbir@122.172.29.116 1300799571 J * BenG ~bengreen@cpc12-aztw24-2-0-cust146.aztw.cable.virginmedia.com 1300799866 N * ensc Guest1751 1300799876 J * ensc ~irc-ensc@p5DF2C4C0.dip.t-dialin.net 1300800000 Q * BenG Quit: I Leave 1300800035 Q * Guest1751 Ping timeout: 480 seconds 1300801256 N * Bertl_zZ Bertl 1300801269 M * Bertl morning folks! 1300801280 M * daniel_hozac morning Bertl 1300801601 N * ensc Guest1756 1300801611 J * ensc ~irc-ensc@p5DF2C4C0.dip.t-dialin.net 1300801764 M * Bertl petzsch: go ahead, test it out, let us know how it goes ... 1300801770 Q * Guest1756 Ping timeout: 480 seconds 1300801996 M * petzsch Bertl: maybe in a week or two when i'll have a new testmachine to play with. actually hoped someone maybe allready gave it a try. 1300802031 M * Bertl maybe someone has, it's around for a while now 1300802114 N * ensc Guest1759 1300802124 J * ensc ~irc-ensc@p5DF2C4C0.dip.t-dialin.net 1300802285 Q * Guest1759 Ping timeout: 480 seconds 1300802298 J * alpha_one_x86 ~kvirc@90.163.39.154 1300802330 M * alpha_one_x86 Hello, what is the main difference between vs2.3.0.36.29 and vs2.3.0.37 ? 1300802489 M * Bertl like half a year of development? 1300802547 M * Bertl (besides the fact that there is no vs2.3.0.37 yet :) 1300802610 M * alpha_one_x86 yes... I plan switch kernel 2.6.32 -> 2.6.38+ in next 6 month 1300802630 M * uranus Bertl, what has to be tested till 2.3.0.37 is released? 1300802761 M * Bertl mainly cleanup and fixing a few minor issues, but of course, cgroup testing and maybe network namespaces as well 1300802926 J * petzsch1 ~markus@p57B6577D.dip.t-dialin.net 1300803220 Q * petzsch Ping timeout: 480 seconds 1300804853 P * bwh 1300804961 J * dowdle ~dowdle@scott.coe.montana.edu 1300805251 N * ensc Guest1767 1300805261 J * ensc ~irc-ensc@p5DF2C4C0.dip.t-dialin.net 1300805420 Q * Guest1767 Ping timeout: 480 seconds 1300807407 Q * petzsch1 Quit: Leaving. 1300807519 M * alpha_one_x86 I can't compile vserver-util: /usr/diet/lib-x86_64/libc.a(sscanf.o): In function `sscanf': (.text+0x1f): undefined reference to `__builtin_stdarg_start' 1300807574 M * alpha_one_x86 Bug with util-vserver-0.30.216_pre2935 but not with: sys-cluster/util-vserver-0.30.216_pre2910 1300807858 M * daniel_hozac that's not a bug in util-vserver. 1300807866 M * daniel_hozac that's diet or your toolchain. 1300807927 J * petzsch ~markus@p57B6577D.dip.t-dialin.net 1300808208 Q * bsingh Ping timeout: 480 seconds 1300808412 M * Bertl alpha_one_x86: on a sidenote: any reason for not using pre2955? 1300808465 M * alpha_one_x86 2955 not present into gentoo portage tree packages 1300808984 J * bsingh ~balbir@122.172.199.4 1300809302 N * ensc Guest1776 1300809312 J * ensc ~irc-ensc@p5DF2C4C0.dip.t-dialin.net 1300809470 Q * Guest1776 Ping timeout: 480 seconds 1300809815 N * ensc Guest1779 1300809825 J * ensc ~irc-ensc@p5DF2C4C0.dip.t-dialin.net 1300809985 Q * Guest1779 Ping timeout: 480 seconds 1300810273 Q * ncopa Quit: Leaving 1300813607 J * bonbons ~bonbons@2001:960:7ab:0:88:c22e:91f5:ff4e 1300817668 Q * petzsch Quit: Leaving. 1300820975 Q * derjohn_mob Ping timeout: 480 seconds 1300821037 J * cluk ~cluk@p4FEA5A66.dip.t-dialin.net 1300821227 M * cluk Hi 1300821260 Q * alpha_one_x86 Quit: KVIrc KVIrc Equilibrium 4.1.1, revision: 5484, sources date: 20110110, built on: 2011-03-07 00:40:35 UTC http://www.kvirc.net/ 1300821319 M * cluk I have setup an ipsec host <-> network tunnel on my host system. Does anybody know if it is possible to use this tunnel from inside a guest? 1300821401 M * Bertl without network namespaces, you can use any IP assigned on the host inside a guest 1300821423 M * Bertl with network namespaces, you should be able to assign any interface to a guest as well 1300821556 N * ensc Guest1802 1300821566 J * ensc ~irc-ensc@p5DF2C4C0.dip.t-dialin.net 1300821571 M * cluk Bertl: Tried to setup the tunnel ip as a second interface on the guest, but it did not seem to work. 1300821607 M * daniel_hozac depending on your settings, you may be using the first guest IP as the source. have you checked your packets as they're leaving? 1300821658 M * cluk daniel_hozac, I have tried: ping -I 1300821681 M * Bertl and that works on the host, I presume? 1300821722 Q * Guest1802 Ping timeout: 480 seconds 1300821802 M * cluk hm. It works without the -I option on the host. But I have to admit, it does not work with the -I option on the host. 1300821818 M * Bertl so there is your problem I guess 1300821832 M * Bertl i.e. it uses a different IP on the host, which is not available to the guest 1300821846 M * Bertl (check with tcpdump -vvnei on the host) 1300821878 J * nkukard ~nkukard@41-133-112-179.dsl.mweb.co.za 1300822070 N * ensc Guest1804 1300822079 J * ensc ~irc-ensc@p5DF2C4C0.dip.t-dialin.net 1300822121 M * cluk ping without -I is using the host's primary ip on the host and the guest's primary ip on the guest. 1300822121 M * cluk as one would expect. :) 1300822144 M * Bertl yep, and your tunnel 'just' works with the primary host IP 1300822238 M * cluk should I be able to use the primary host ip in a guest? 1300822240 Q * Guest1804 Ping timeout: 480 seconds 1300822251 J * yarihm ~yarihm@guest-docking-nat-1-034.ethz.ch 1300822253 M * Bertl unless you assign it to the guest, no 1300822300 M * Bertl i.e. what I'm saying is, either your tunnel is badly configured or you want your guest to use the host IP :) 1300822658 M * cluk hm. As I am unable to configure the remote side of the tunnel, I have now added the host's primary ip to the guest. 1300822697 M * cluk using this ip with: ping -I from inside the host still does not work. 1300822749 M * cluk neither does the same command from the host. 1300822770 M * cluk the only working version is a ping without the -I from the host 1300822787 J * manana ~mayday090@84.17.25.149 1300822816 M * Chlorek what tcpdump says about it? 1300822836 M * Bertl cluk: so, you are telling me that ping (without -I) uses the host IP 1300822839 M * Chlorek is this ip a local? 1300822846 M * Chlorek behind nat 1300822854 M * Bertl and using ping -I fails even on the host? 1300822875 M * Bertl if so, I would strongly recommend upgrading to a working kernel :) 1300822898 M * cluk :) 1300822936 M * cluk ok, found it. 1300823004 M * cluk I simply swapped the local and remote tunnel ip in my first attempt. 1300823012 M * Bertl okay :) 1300823034 M * Bertl for the assign host IP to the guest part, I would consider using S/DNAT instead 1300823039 M * cluk adding the local tunnel ip to the guest and using this with ping -I works on the host and the guest. 1300823057 M * Bertl i.e. assign a private IP to the guest, and then S/DNAT that IP to the host's tunnel IP 1300823098 M * cluk thanks a lot for your help. 1300823110 M * Bertl mainly because it adds some flexibility regarding changes on the tunnel IP as well as avoids colissions between host and guest 1300823115 M * Bertl you're welcome! 1300823193 N * ensc Guest1806 1300823203 J * ensc ~irc-ensc@p5DF2C4C0.dip.t-dialin.net 1300823247 J * derjohn_mob aj@88.128.41.33 1300823362 Q * Guest1806 Ping timeout: 480 seconds 1300823406 M * cluk Bertl, the SNAT approach had been my very first attempt. but it did not work using the swapped tunnel ip. ;) 1300823423 M * cluk Working fine with SNAT and without an additional guest ip now. 1300823432 M * cluk Thanks again. 1300823569 M * Bertl np 1300825282 Q * derjohn_mob Ping timeout: 480 seconds 1300825693 J * Piet_ ~Piet__@28IAAAJ9D.tor-irc.dnsbl.oftc.net 1300825882 Q * cluk Quit: Ex-Chat 1300825937 Q * Piet Ping timeout: 480 seconds 1300826727 Q * ghislain Quit: Leaving. 1300827001 J * pinochle ~pinochle@64.7.104.27 1300827067 M * pinochle hi guys, Im migrating a vserver from a different box and I can't get it to start 1300827069 M * pinochle root@tequila:/etc/vservers/absinthe# vserver absinthe start 1300827069 M * pinochle vcontext: execvp("/etc/init.d/rc"): No such file or directo 1300827140 M * pinochle /etc/init.d/rc is present in the vserver root 1300827793 Q * manana Remote host closed the connection 1300828069 J * derjohn_mob aj@88.128.41.150 1300828329 P * pinochle 1300828354 J * petzsch ~markus@p57B6577D.dip.t-dialin.net 1300828367 M * Bertl hmm ... just missed him/her 1300829080 Q * derjohn_mob Ping timeout: 480 seconds 1300829091 J * derjohn_mob aj@88.128.41.150 1300829173 M * daniel_hozac oh well 1300829512 Q * yarihm Quit: This computer has gone to sleep 1300829641 J * BenG ~bengreen@cpc12-aztw24-2-0-cust146.aztw.cable.virginmedia.com 1300830485 J * pinochle ~pinochle@64.7.104.27 1300830494 M * pinochle anyone around atm? 1300830508 M * daniel_hozac yep 1300830520 M * pinochle oh good :D 1300830578 M * pinochle im trying to migrate a debian-etch guest to a new squeeze host, but gettings an error it cant find /etc/init.d/rc 1300830618 M * daniel_hozac how did you migrate it? 1300830790 M * pinochle just copied it 1300830810 J * edit ~blagoj@89.205.105.228 1300830831 M * edit hi 1300830857 M * pinochle daniel_hozac: I made a tar of the old guest root, and used the template build option 1300830875 M * edit can someone please explain me what values in "/etc/vservers//schedule" means? 1300830909 M * edit I curentely have these and couple of VPS-s that I have are working just fine. 1300830911 M * edit cat /etc/vservers//schedule 1300830911 M * edit 14 1300830911 M * edit 100 1300830911 M * edit 0 1300830911 M * edit 10 1300830912 M * edit 400 1300830912 M * edit 0 1300830923 M * edit But I wint to know what they mean? 1300830924 M * pinochle http://oldwiki.linux-vserver.org/Scheduler+Parameters 1300830935 M * pinochle edit: http://oldwiki.linux-vserver.org/Scheduler+Parameters 1300830937 M * pinochle soorry 1300831123 M * edit Thanks, so from what can I understand this is a "kind" of how the processor will be use/split betvean the VPS's, right? 1300831137 M * pinochle yes, cpu schedulintg 1300831146 M * pinochle think you need a kernel patch 1300831169 M * edit What kernel patch? 1300831289 M * Bertl it's part of older Linux-VServer 1300831311 M * daniel_hozac pinochle: rsync probably would've worked better. can you run vserver start --rescue and vserver enter in another shell? 1300831313 M * Bertl it was replaced in recent kernels with the cgroup way of handling scheduling limits 1300831372 M * Bertl pinochle: you can still (re)run rsync with numeric ids on the directory if the old guest root is intact 1300831386 M * edit I know but I'm curentely experimenting with 2.6.33.7-vs2.3.0.36.30.4 kernel, I have 2.6.36.2-vs2.3.0.36.38.2 on other machine and I will play with it some of the next incoming days. 1300831398 M * edit I first want to see the "history" :) 1300831401 M * pinochle Bertl / daniel_hozac : daniel_hozac: is that right way 1300831403 M * pinochle oops 1300831411 M * pinochle Bertl / daniel_hozac : root@tequila:~# vserver absinthe start --rescue 1300831411 M * pinochle vcontext: execvp("/bin/sleep"): No such file or directory 1300831447 Q * BenG Quit: I Leave 1300831452 M * Bertl edit: both your kernels do not have the hard cpu scheduler anymore (i.e. you want to use cgroups anyway) 1300831472 M * daniel_hozac pinochle: does chroot /vservers/ work? 1300831478 M * edit Sure, thanks. 1300831481 M * pinochle root@tequila:~# ls -alh /var/lib/vservers/absinthe/bin/sleep 1300831481 M * pinochle -rwxr-xr-x 1 root root 30K Mar 22 14:38 /var/lib/vservers/absinthe/bin/sleep 1300831489 M * edit Can you please tell me what was the page where everything what resides in /etv/vservers/ was explained? 1300831490 M * Bertl pinochle: either you are missing part of the guest libraries, or you have a problem with the architecture 1300831517 M * Bertl as daniel_hozac suggests, try something like chroot /path/to/guest 1300831531 M * pinochle root@tequila:~# chroot /var/lib/vservers/absinthe/ 1300831531 M * pinochle chroot: failed to run command `/bin/bash': No such file or director 1300831540 M * pinochle wtf... 1300831543 M * pinochle thats weird 1300831545 M * edit Nevermind, I find it. 1300831546 M * edit http://www.nongnu.org/util-vserver/doc/conf/configuration.html 1300831557 M * daniel_hozac google "great flower page" 1300831569 M * Bertl pinochle: see, now the question is, is it an architecture issue (e.g. 64bit vs 32bit) or a binary/library issue 1300831580 Q * bonbons Quit: Leaving 1300831591 M * pinochle Bertl: I can run the binaries in the vserver from the host 1300831599 M * daniel_hozac pinochle: so probably libs. 1300831615 M * Bertl then most likely some libraries are missing or have wrong flags/permissions 1300831617 M * pinochle for example root@tequila:~# /var/lib/vservers/absinthe/bin/sleep 1 1300831620 M * pinochle works 1300831654 M * pinochle poop 1300831662 M * pinochle how do i fix the libraries? 1300831750 M * pinochle ok ill try rsyncing from the backup 1300831835 M * pinochle daniel_hozac / Bertl: if that doesnt work, am I just screwed? 1300831859 M * daniel_hozac no... 1300831867 M * pinochle root@tequila:~# vserver absinthe start 1300831867 M * pinochle vcontext: execvp("/etc/init.d/rc"): No such file or directory 1300831884 M * pinochle same problem after rsync... weird because all the permissions seem the same 1300831937 M * Bertl rsync with what options? 1300831989 M * pinochle rsync -rv /root/migration/var/lib/vservers/absinthe /var/lib/vservers/ 1300831999 M * Bertl try -axHPSD --numeric-ids :) 1300832021 M * daniel_hozac where did /root/migration come from? 1300832066 M * daniel_hozac that doesn't look like it's the original data. 1300832097 M * pinochle i tar -cvf 'd the original data and then scp'd and extracted it to the new host 1300832105 M * pinochle daniel_hozac: so its not really 1300832114 M * daniel_hozac copying from an incorrect source isn't going to help you. 1300832146 M * Bertl i.e. you want to rsync _from_the_original_ :) 1300832160 M * pinochle okay... is rsync over ssh o.k.? 1300832165 M * Bertl yep 1300832384 M * pinochle okay ill try again reboot required ;) 1300832388 M * pinochle brb 1300832409 Q * pinochle Quit: Leaving. 1300833564 J * d0ne ~coma@c-69-181-70-211.hsd1.ca.comcast.net 1300833769 M * d0ne hi, I was wondering if someone could help me with vserver and dhcp ? 1300833806 M * daniel_hozac depends on what you're trying to do. 1300833811 M * d0ne I'm trying to get dhclient work in a vserver guest 1300833818 M * d0ne but that doesn't seem to work as expected 1300833852 M * Bertl do you use network namespaces with an ethernet device? 1300833867 M * Bertl (note that's not the default for Linux-VServer :) 1300833888 M * d0ne i just built a default vserver to test with with --interface eth0..... 1300833902 M * d0ne added the bcapabilities flags and then tried to run dhclient inside 1300833904 M * Bertl so in this case, you are using IP isolation 1300833921 M * d0ne I couldn't figure out at all how to have dhcp from the moment you start the guest 1300833922 M * Bertl and you don't want to use dhcp inside the guest 1300833975 M * d0ne i read about IPROOT=dhcp, but couldn't figure out how to use this config flag 1300833998 M * Bertl where did you read about that? 1300834015 M * daniel_hozac jacques tools :) 1300834039 M * d0ne http://tycoontalk.freelancer.com/website-and-server-administration-forum/127989-using-dhcp-on-servers.html 1300834058 M * d0ne but yeah as said didn't work out :) 1300834098 M * Bertl the thing is, even with IP isolation, you can simply 'request' another IP from your dhcp server (on the host) 1300834101 M * d0ne is there a possibility to have dhcp when you start the guest i.e. vserver test start and then it will get an ip from dhcp ? 1300834114 M * Bertl and use that for the guest, i.e. assign that IP to the interface and guest 1300834130 M * Bertl there are startup and shutdown scripts for each guest, where you can do that 1300834174 M * Bertl (but check first, that your dhcp server hands out IPs based on client identifier) 1300834182 M * d0ne where could I find an example of such a startup script ? 1300834202 M * d0ne the dhcpd does, cause it was setup exactly for this test :) 1300834337 M * daniel_hozac just run another dhclient on the host. 1300834353 M * Bertl I should have an example somewhere 1300835311 N * Piet_ Piet 1300835884 Q * FireEgl Ping timeout: 480 seconds 1300835906 J * FireEgl ~FireEgl@173-25-19-139.client.mchsi.com 1300836122 Q * petzsch Quit: Leaving. 1300836618 Q * dowdle Remote host closed the connection