1138665605 M * daniel_hozac $3? won't that get the prefix? 1138665611 M * Bertl or shorter: 1138665612 M * Bertl ip addr ls eth0 | gawk -F "[ /]*" '/inet / { print $3 }' 1138665645 M * Bertl daniel_hozac: hmm, nope :) 1138665652 M * daniel_hozac i see that, but why? 1138665665 M * Bertl because gawk is strange with FS 1138665675 M * Bertl you get an empty string at the beginning 1138665693 M * daniel_hozac oh, ok. 1138665698 J * ScoobyD00 ~foo@80-195-186-201.cable.ubr08.newm.blueyonder.co.uk 1138665907 Q * miRC Quit: I'm seeing someone else now Jerake, sorry - http://www.lmfao.tk/ 1138665926 M * ScoobyD00 can the latest patch go against 2.6.8? :o) 1138665942 M * Bertl ScoobyD00: well, if you abckport it, yes :) 1138665948 M * Bertl *backport 1138665989 M * ScoobyD00 ;o) 1138666021 M * ScoobyD00 still trying to figure out how to get the software more up to date on my debian install without opening the floodgates and losing security updates 1138666046 M * ScoobyD00 what kernels does 2.1.0 support - just 2.6.14? 1138666081 M * Bertl 2.1.x is devel, you probably want 2.0.1 ... 1138666098 M * ScoobyD00 doh - yeah on wrong page :o) 1138666112 M * undefined ScoobyD00: i'm using ubuntu's 2.6.12 on debian patched with 2.0 1138666118 M * Bertl ScoobyD00: but regarding 'losing security updates', what about a newer debian kernel? 1138666141 M * ScoobyD00 well, there are no newer kernels in "stable" which is the area that gets security updates as i understand 1138666158 M * ScoobyD00 undefined: i have been thinking about tracking my kernel off ubuntu too :o) 1138666177 M * undefined the ubuntu security mailing list is low volume 1138666191 M * Bertl ScoobyD00: well, mainline is tracking security updates pretty closely too, no? 1138666191 M * undefined ScoobyD00: easy to spot kernel updates on it 1138666207 M * ScoobyD00 undefined: i couldnt find 2.6.12 on packages.ubuntu.com, only 2.6.11 1138666224 M * undefined ScoobyD00: let me find it for you, it's there 1138666230 M * undefined ScoobyD00: they just had a security update last week 1138666248 M * undefined ScoobyD00: were you searching on the right distro 1138666263 M * ScoobyD00 undefined: oh, it got renamed, right? 1138666271 M * ScoobyD00 from kernel-source to linux-source 1138666278 M * undefined ScoobyD00: http://packages.ubuntu.com/cgi-bin/search_packages.pl?keywords=linux-source&searchon=names&subword=1&version=all&release=main 1138666298 M * undefined ScoobyD00: the key is "linux-source" "all" and "main" 1138666328 M * ScoobyD00 bertl: will 2.0.1 work against 2.6.12? :o) 1138666362 M * undefined ScoobyD00: i've applied the patches from http://vserver.13thfloor.at/Devel/PAT-2.0.1/ to ubuntu's 2.6.12 1138666375 M * Bertl ScoobyD00: no, but 2.0 will 1138666377 M * undefined ScoobyD00: (after applying 2.0) 1138666381 M * ScoobyD00 undefined: fab :0) 1138666421 M * undefined ScoobyD00: i've applied all those patches except one where the kernel appears to have broken an inet stack out from tcp (or something like that) 1138666424 M * Skram has anyone made a good php/perl control panel for vservers to restart them? 1138666446 M * Skram I want to give some access to start and shutdown vservers, but i dont want them to have root access to the vpshot 1138666468 M * undefined ScoobyD00: the patch doesn't apply because the file does not exist in 2.6.12 (but in 2.6.14 which 2.0.1 targets) 1138666474 M * Bertl Skram: what about a 'special' ssh account? 1138666480 M * Skram Bertl: possible.. 1138666495 M * Skram but they would have access to "vserver" so they can do vserver enter 1138666501 M * Skram I havent messed with shells 1138666504 M * Skram always use bash :( 1138666517 M * Bertl Skram: no, just for the restart command 1138666523 M * Skram hmm 1138666525 M * Skram how? 1138666586 M * Bertl http://www.sculpturedlife.com/vserver/irc-logs/index.php?date=2004-07-08 1138666612 M * Bertl search for trampoline 1138666625 M * Skram hmm okay 1138666655 M * Skram http://vserver.13thfloor.at/Stuff/trampoline.sh 1138666656 M * ScoobyD00 undefined: are the headers supposed to be at: linux-kernel-headers 2.6.11.2-0ubuntu13 1138666659 M * Skram Object Not Found 1138666676 M * Bertl Skram: well, it moved since, but it's not required for your case 1138666720 M * Bertl http://vserver.13thfloor.at/Stuff/SCRIPT/trampoline.sh 1138666751 M * undefined ScoobyD00: just curious, what are you needing the headers for? 1138666763 M * ScoobyD00 well, it pulled them in without asking 1138666783 M * undefined ScoobyD00: dependency on linux-source? 1138666783 M * Skram im a little confused 1138666786 M * Skram ill look into it 1138666839 M * ScoobyD00 undefined: yeah - im not so up with linux (more used to FreeBSD), so if you say i done need them, thats fine :o) 1138666847 M * Bertl Skram: the idea is to make a special script to reboot a guest, then get the customers key, and add it (or build the key for him, and send it to him) 1138666972 M * undefined ScoobyD00: i usually grab linux-source-2.6.12 (for the source) and linux-headers-2.6.12-10-amd64-k8 (for the config) 1138666998 Q * marl_ Read error: Connection reset by peer 1138667008 J * marl ~matt@84.92.193.226 1138667030 M * undefined ScoobyD00: i only use the config file from linux-headers-2.6.12-10-amd64-k8 and throw away everything else 1138667048 M * undefined ScoobyD00: you don't have to use the same config ubuntu used to compiler their kernel, but "when in rome, do as the romans" 1138667155 M * undefined ScoobyD00: then (to keep this ontopic ;)) i apply the vs 2.0 patch 1138667790 J * marl_ ~matt@84.92.193.226 1138667790 Q * marl Read error: Connection reset by peer 1138667925 Q * flock Ping timeout: 480 seconds 1138668131 Q * mountie Remote host closed the connection 1138668399 J * mountie ~mountie@CPEdeaddeaddead-CM000a739acaa4.cpe.net.cable.rogers.com 1138668496 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il 1138668612 M * sladen ScoobyD00: packages.ubuntu.com doesn't have dapper. 2.6.12 is dapper 1138668628 M * sladen urm, 2.6.15 1138668640 M * sladen *tired*, night night all 1138668931 M * Bertl night sladen! 1138669511 P * undefined 1138670312 M * mugwump cool, now I can import those quilt patchsets into git with one command! :) 1138670355 M * mugwump is there a nopaste in the house? 1138670385 M * Bertl hmm? 1138670397 M * mugwump er, somewhere I can nopaste to ... I'll just use rafb.net 1138670411 M * Bertl make that 1138670513 M * phycho hey guys got a problem with util-vserver 1138670514 M * phycho checking for C++ compiler default output file name... configure: error: C++ compiler cannot create executables 1138670514 M * phycho See `config.log' for more details. 1138670516 M * phycho anyone got any ideas? 1138670520 M * phycho thats after running ./configure 1138670534 M * phycho the log just states 1138670534 M * phycho configure: exit 77 1138670555 M * mugwump phycho: what happens when you run the command in the log immediately above that error? 1138670585 M * phycho sec 1138670588 M * phycho i think i see the problem 1138670808 M * phycho Can not find the 'vconfig' tool within '/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/bin/X11:/sbin:/usr/sbin:/usr/local/sbin'. 1138670808 M * phycho See http://www.candelatech.com/~greear/vlan.html; usually this tool is shipped in the 'vconfig' or 'vlan' package of your distribution 1138670848 M * Doener no, that's unrelated 1138670864 M * phycho yeah i fixed the other one 1138670932 M * Doener the vconfig one just tells that you don't have vlan tools installed and thus util-vserver won't support them 1138670947 M * Doener if you don't need vlan support, ignore it 1138670948 M * phycho k 1138671042 Q * marl_ Read error: Connection reset by peer 1138671048 M * Doener hm, it's required now? i thought it was optional... oops, didn't compile util-vserver myself for too long abviously 1138671054 M * phycho its required 1138671055 M * Doener s/abviou/obviou/ 1138671071 J * marl ~matt@84.92.193.226 1138671088 M * phycho :p 1138671277 A * phycho wonders if he should create a debian package out of it 1138671702 Q * ScoobyD00 Quit: 1138672134 M * mugwump ~http://rafb.net/paste/results/LGgcSy34.html # integration patch plan 1138672141 M * mugwump s/~// 1138672262 M * Bertl interesting ... 1138672511 M * phycho Filesystem Size Used Avail Use% Mounted on 1138672511 M * phycho /dev/hda1 221M 212M 0 100% / 1138672512 M * phycho oh oh... 1138672625 M * mugwump Bertl: actually I think only the early patches will require much work 1138672639 M * mugwump (where "work" is basically splitting and making each point consistent) 1138672934 Q * lilalinux Remote host closed the connection 1138673432 M * Bertl mugwump: well, let me know if you need anything (e.g. an updated split or so) 1138673502 M * phycho hey guys 1138673504 M * phycho ive got a problem 1138673509 M * phycho uml:/usr/local/var/run/vservers# vserver www start 1138673509 M * phycho /proc/uptime can not be accessed. Usually, this is caused by 1138673509 M * phycho procfs-security. Please read the FAQ for more details 1138673509 M * phycho http://www.linux-vserver.org/index.php?page=Linux-Vserver+FAQ 1138673509 M * phycho :X 1138673568 M * Bertl /etc/init.d/vprocunhide start 1138673588 M * phycho no such file/directory 1138673602 M * Bertl well, or wherever your distro puts it 1138673619 M * Bertl (that's the path where the utils install when compiled from scratch) 1138673633 M * phycho vie found it.. 1138673645 M * phycho but it hangs when run 1138673682 M * phycho and it eats the cpu 1138673687 M * phycho it supposed to do that? 1138673720 M * Bertl no, usually it doesn't do that 1138673747 M * phycho well it is on my box 1138673748 M * phycho Cpu(s): 96.7% us, 3.3% sy, 0.0% ni, 0.0% id, 0.0% wa, 0.0% hi, 0.0% si 1138673749 M * phycho ;/ 1138673760 M * phycho any ideas? 1138673780 M * Bertl strace 1138673792 M * Bertl to see what it does ... 1138673809 M * phycho that wont help me any 1138673813 M * phycho i dont understand a word of it 1138673835 M * Bertl what was your tool and kernel version? 1138673843 M * phycho latest tool 1138673849 M * phycho Linux uml 2.6.14.6 #1 Mon Jan 30 00:38:05 GMT 2006 i686 GNU/Linux 1138673849 M * phycho uml:/usr/local/lib/util-vserver# 1138673851 M * phycho thats kernel 1138673883 M * Bertl very strange ... 1138673904 M * Bertl it's uml arch, okay, but vserver patched? 1138673924 M * phycho eh.. its not uml at all 1138673928 M * phycho thats just the name of the box 1138673935 M * phycho its vserver patched, yes 1138673936 M * Bertl ah, okay, sorry 1138673946 M * phycho uml:/usr/local/lib/util-vserver# vserver www start 1138673946 M * phycho vserver 'www' already running 1138673948 M * Bertl the testme.sh was successful? 1138673948 M * phycho it does that too 1138673952 M * phycho what testme.sh? 1138673962 M * phycho there wasnt any 1138673972 M * Bertl sec, it's in the howto 1138673980 M * phycho its not... 1138673985 M * phycho http://www.5dollarwhitebox.org/wiki/index.php/Howtos_Linux-Vserver_Debian_Sarge 1138673990 M * Bertl http://vserver.13thfloor.at/Stuff/SCRIPT/testme.sh-0.15 1138674004 M * Bertl give it a whirl on the host (as root) 1138674014 M * Bertl upload the output to pastebin.com 1138674102 M * phycho it all suceeded 1138674104 M * phycho k 1138674126 M * phycho http://pastebin.com/531502 1138674158 M * Bertl okay, looks goo ... 1138674160 M * Bertl *good 1138674173 M * Bertl was the guest running when you started it? 1138674179 M * phycho nope 1138674185 M * phycho i can run it and retest it if u want 1138674186 M * phycho ? 1138674194 M * Bertl check if something is running now with 'vserver-stat' 1138674204 M * phycho nope 1138674213 M * phycho if i run 'www' (the guest i created) it loads and is in there now 1138674234 M * Bertl run? 1138674242 M * phycho uml:~# vserver www start 1138674249 M * Bertl okay, so it starts now? 1138674253 M * phycho uml:~# vserver-stat 1138674256 M * phycho 49158 5 11M 2.7M 0m00s10 0m00s00 0m10s24 www 1138674256 M * phycho uml:~# 1138674257 M * phycho yep 1138674259 M * phycho but no networking... 1138674271 M * Bertl well, sure about that? 1138674275 M * phycho yep 1138674280 M * Bertl how do you check? 1138674283 M * phycho if i go into it i get 1138674289 M * phycho www:/# ifconfig 1138674289 M * phycho eth0 Link encap:Ethernet HWaddr 00:40:F4:74:54:89 1138674295 M * phycho no ip assigned to interface... 1138674305 M * Bertl don't use ifconfig 1138674311 M * phycho www:/# ifconfig eth0 192.168.0.1 1138674311 M * phycho SIOCSIFADDR: Permission denied 1138674315 M * Bertl (unless you configured an alias) 1138674324 M * phycho i didnt 1138674325 M * Bertl use 'ip addr ls' 1138674334 M * phycho no ip 1138674338 M * phycho (no ip command) 1138674345 M * Bertl well, install it :) 1138674347 M * mnemoc install iproute2 1138674362 M * phycho k 1138674366 M * mnemoc ifconfig got absolete about 7 years ago 1138674377 M * phycho hehe 1138674391 M * Bertl yes, that's the problem 1138674400 M * Bertl but you can work around by using the older aliases 1138674419 M * phycho inet 192.168.0.7/24 brd 192.168.0.255 scope global eth0 1138674463 M * phycho how come if i ssh to it, i get the host users ssh? 1138674487 M * phycho is it because it can only listen on an ip and not bind to all interfaces? 1138674554 M * Bertl no, that's because your host sshd is already bound to 0.0.0.0 ... a Listen directive and restart will help there 1138674580 M * phycho aah 1138674582 M * phycho k np 1138674624 M * phycho any quick way of changing the ips it uses? 1138674631 M * phycho in the vserver 1138674654 M * Bertl the vserver will always use the assigned ips 1138674664 M * phycho k 1138674665 M * Bertl and the 0.0.0.0 binding will also be restricted to that 1138674673 M * Bertl the host's sshd is the problem 1138674685 M * phycho yeh 1138674697 M * Bertl the listen directive is the solution :) 1138674697 M * phycho will fix it in a mi 1138674698 M * phycho *min 1138674704 M * phycho yep 1138674711 M * Bertl but, as an alternative, you can use a different port inside the guest 1138674719 M * phycho yeh 1138674769 M * phycho can u firewall between them? 1138674844 A * phycho thinks about setting up vlans for them 1138674886 M * Bertl you can do everything as on a normal host 1138674890 M * phycho btw vserver is nice =) 1138674899 M * Bertl just keep in mind that _all_ networking happens on the host 1138674901 M * phycho much easier than user mode linux. 1138674901 M * Skram named will work INSIDE a vps right? 1138674915 M * phycho i would guess so 1138674918 M * Bertl i.e. there is (for now) no virtual network or swittch or whatever 1138674921 M * phycho since it worked on my lycos vds 1138674931 M * phycho and it was based on vserver 1138674954 M * Bertl yes, the lycos vds are an older version of linux-vserver 1138674955 Q * lonewolff Ping timeout: 480 seconds 1138674968 M * phycho bertl - what happens if a user on a vds binds to all ports? 1138674974 M * phycho does it screw up the system? 1138674980 M * phycho aka does it take over all the other ports? 1138674989 M * Bertl no 1138674990 M * phycho on all the servers? 1138674997 M * phycho for example 1138675000 M * phycho if the host had 192.168.1.2 1138675005 M * phycho and the guest had 192.168.1.3 1138675007 M * Bertl simply because each guest is restricted to the assigned ips 1138675011 M * phycho ah 1138675013 M * phycho yeh thats what i meant 1138675014 M * phycho heh 1138675021 M * Bertl only the host is not restricted 1138675027 M * phycho couldnt the user in the guest system change it tho? 1138675031 M * phycho (change his ip allocation?) 1138675033 M * Bertl no 1138675036 M * phycho k 1138675037 M * phycho sweet 1138675040 M * Bertl as you saw with ifconfig 1138675046 M * phycho heh yeah 1138675050 M * Bertl assigning an ip inside is not possible 1138675053 M * phycho k 1138675053 M * phycho cooo 1138675055 M * phycho *cool 1138675075 M * phycho any way to make it show up properly in ifconfig? 1138675087 M * Bertl yes, the beforementioned alias 1138675092 M * phycho k 1138675099 A * phycho scrolls up 1138675118 M * Bertl on creation, just use --interface karli=eth0:192.168.2.1/24 1138675134 M * Bertl or afterwards, stop the guest, and add a 'name' file to the config 1138675147 M * Bertl /etc/vservers//interfaces/0/name 1138675163 M * Bertl which contains the alias extension 1138675178 M * phycho k :) 1138675216 Q * michal_ Ping timeout: 481 seconds 1138675372 M * phycho what does rebootmgr do? 1138675447 M * Bertl nothing anymore, was replaced by vshelper 1138675452 J * lonewolff ~lonewolff@host-84-9-143-159.bulldogdsl.com 1138675459 M * phycho k 1138675476 M * phycho any way i can make the vservers start on boot? 1138675494 M * phycho (wheres the default boot script?) 1138675518 M * Bertl yes, there is a script (like the vprocunhide) which should be started on system startup 1138675528 M * Skram named: capset failed: Operation not permitted: please ensure that the capset kernel module is loaded. see insmod(8) 1138675534 M * Skram hercules / # modprobe -l | grep CONFIG_SECURITY_CAPABILITIES 1138675534 M * Skram hercules / # 1138675542 M * Skram no module or setting is present, hrmph? 1138675545 M * Bertl which in turn, starts all 'marked' guests with 'default' in the mark file (check Flower Page for the path 1138675546 M * phycho bertl - can you tell me the name of it please? 1138675554 M * phycho k 1138675557 M * phycho np 1138675587 J * michal_ ~michal@www.rsbac.org 1138675590 M * phycho vservers-default 1138675593 M * phycho thats it i guess? 1138675600 M * phycho Starting vservers of type 'default'.... 1138675602 M * Bertl Skram: two options here: get a new bind (without the caps enabled), or, give potentially dangerous caps to the guest 1138675615 M * Bertl Skram: see FAQ page and problematic programs 1138675698 M * Skram So its a vserver thing? 1138675700 M * Skram not a kernel thing 1138675701 M * Skram # CONFIG_SECURITY is not set 1138675703 M * Skram :| 1138675711 M * Bertl phycho: exactly 1138675724 M * Bertl Skram: hmm, how did that happen? 1138675728 M * Skram Not sure 1138675733 M * Skram I didnt setup this system 1138675743 M * Bertl well, the linux-vserver patches should auto-select that 1138675754 M * Skram weird 1138675773 M * phycho eer my eyes 1138675780 A * phycho screams at flower page 1138675825 M * Bertl phycho: 'weedpage' 1138675850 M * phycho is it the context variable i need? 1138675856 M * phycho to start 'default' servers? 1138675879 M * Bertl no, that#s the stylesheet I'd suggest :) 1138675882 A * phycho is reading page 1138675883 M * phycho heh k 1138675895 M * Bertl look for 'mark' as I said 1138675900 M * phycho aah. 1138675900 M * phycho mark 1138675902 M * phycho k mp 1138675904 M * phycho *np 1138675907 A * phycho is half asleep 1138675909 M * phycho 3am here ;) 1138675946 M * phycho oooh.. i didnt know vserver did resource limits 1138675948 M * phycho sweet@ 1138675949 M * phycho *! 1138675957 A * phycho bookmarks page 1138675988 M * Bertl 4am here :) 1138676073 M * phycho hehe 1138676119 M * phycho vserver is damned sweet =))) 1138676149 M * Bertl indeed it is .. btw, here is a feature matrix: 1138676152 M * phycho shame it doesnt start the damned thing :@ 1138676161 M * phycho i added that default mark file 1138676171 M * phycho and ran vservers-default 1138676174 M * phycho and it didnt start it up 1138676211 M * phycho uml:/usr/local/etc/vservers/www/apps# 1138676214 M * phycho mark is in there 1138676218 M * phycho uml:/usr/local/etc/vservers/www/apps# cat mark 1138676219 M * phycho default 1138676226 M * phycho and when i run the start default script, its not started. 1138676227 M * Bertl check if it contains a newline 1138676237 M * Bertl 'default\n' 1138676246 M * Bertl http://linux-vserver.org/Release+FAQ 1138676257 M * phycho does it need to? 1138676261 M * phycho atm its just 'default' 1138676287 M * Bertl IIRC, the tools always want a newline 1138676300 M * phycho k 1138676414 M * phycho do i need the ''? 1138676440 M * Doener Bertl: hm, didn't the vserver patch auto-select CONFIG_SECURITY_CAPABILITIES when CONFIG_SECURITY is set, and not auto-select CONFIG_SECURITY? 1138676450 M * Doener or was that changed? 1138676475 M * phycho bertl - ive added it, with the '', it still didnt work 1138676496 M * phycho tried without and that didnt work either. 1138676550 M * Bertl Doener: hmm, hmm, without config security, the default caps should be included 1138676592 M * Bertl phycho: well, it works here, but I'd suggest to run the script with 'bash -x ' to see what it does, and where it fails 1138676606 M * phycho k 1138676629 M * Doener yep, but the bind-with-caps problem remains, or is there a workaround now? 1138676671 M * Skram The only complaint i have with linux-vserver is that users cant have their own kernel, but it is also one of the identifying characters of VServer, so ;( 1138676672 M * Doener phycho: it's apps/init/mark 1138676700 M * Bertl in theory, we could allow the bind resource cap ... 1138676712 M * Skram im enabling it on the host... 1138676721 M * Skram it wont automatically be accessible by vservers? 1138676736 M * Bertl Doener: but it isn't tested so we do not do it for now ... 1138676737 M * phycho aah 1138677767 Q * Doener Quit: Leaving 1138677831 M * phycho http://www.darktech.org.uk/phpsysinfo/ 1138677832 M * phycho that work? 1138678052 M * Bertl seems so 1138678084 M * phycho k cool 1138678085 M * phycho ty 1138678088 M * phycho thats a vserver =) 1138678106 M * Bertl congrats! 1138678111 M * phycho hehe :) 1138678126 M * phycho all i need now is a seperate one for mail services 1138678136 M * Bertl you might want to add yourself to the Happy Users page (when you're done) 1138678139 M * phycho what MTA do you guys suggest? 1138678145 M * Bertl postfix 1138678146 M * phycho =) yeah 1138678148 M * phycho k. 1138678167 M * phycho ive used exim before, but fancy a change 1138678553 A * mugwump returns from a meeting that took far too long 1138678567 M * mugwump Bertl: a new split would be nice 1138678593 M * mugwump just your current, don't bother about updating to latest kernel rev 1138678758 M * Bertl mugwump: okay, np, but tomorrow .. for now, I'm off to bed ... 1138678773 M * Bertl night everyone, have a good whatever .. cya later! 1138678777 M * mugwump seeya 1138678784 N * Bertl Bertl_zZ 1138678794 J * xylianai undead@c220-239-105-68.lowrp1.vic.optusnet.com.au 1138678821 P * xylianai 1138680770 J * Al ~Al@62.38.108.5 1138680791 P * Al 1138681922 J * jkl eric@c-67-172-156-116.hsd1.co.comcast.net 1138683253 Q * phycho Read error: Connection reset by peer 1138683911 J * balbir ~balbir@59.145.136.1 1138684602 J * [^GaLaCtIcO^] ~ircap8@host2.200-82-90.telecom.net.ar 1138684608 P * [^GaLaCtIcO^] 1138684626 J * Rubia ~ircap8@host2.200-82-90.telecom.net.ar 1138686589 Q * Rubia Quit: � IRcap [8.0] � www.ircap.net � 1138688729 Q * _are_ Ping timeout: 480 seconds 1138689949 Q * mnemoc Ping timeout: 480 seconds 1138689958 J * DoM_ ~DoM@151.56.238.63 1138690349 Q * DoM__ Ping timeout: 480 seconds 1138690737 J * mnemoc ~amery@200.75.27.81 1138692229 Q * Hunger arion.oftc.net europa.oftc.net 1138692306 J * Hunger Hunger.hu@Hunger.hu 1138693330 J * Smutje ~Smutje@xdsl-84-44-242-189.netcologne.de 1138693439 Q * Smutje_ Ping timeout: 480 seconds 1138693779 J * meandtheshell ~markus@85-125-228-78.dynamic.xdsl-line.inode.at 1138693987 J * Johnnie ~jdlewis@dynamic-acs-24-154-53-16.zoominternet.net 1138695298 J * _are_ ~are@62.112.159.81 1138697293 Q * aba Ping timeout: 480 seconds 1138697622 Q * _are_ Read error: Connection reset by peer 1138697947 J * ntrs_ ~ntrs@68-188-50-87.dhcp.stls.mo.charter.com 1138697947 Q * ntrs__ Read error: Connection reset by peer 1138698196 J * dothebart ~willi@xdsl-81-173-175-220.netcologne.de 1138698626 Q * tudenbart Ping timeout: 480 seconds 1138698959 J * prae ~prae@ezoffice.mandriva.com 1138699189 Q * Johnnie Remote host closed the connection 1138699458 J * gerrit ~gerrit@visit.suse.de 1138700911 J * Lars ~adminuser@you.are.the.lamer.la 1138700912 M * Lars Hello 1138700919 M * Lars why are my proftpd services keeping stoning? 1138701015 J * menomc ~amery@200.75.27.31 1138701121 Q * mnemoc Ping timeout: 480 seconds 1138701121 N * menomc mnemoc 1138701717 M * SiD3WiNDR stoning? 1138705714 Q * Aiken Ping timeout: 480 seconds 1138707499 M * Hollow olla 1138707596 M * Hollow Skram: ping 1138707941 J * bonbons ~bonbons@83.222.38.150 1138708074 M * Hollow heya bonbons 1138708085 M * bonbons hey Hollow 1138708160 Q * FireEgl Ping timeout: 480 seconds 1138708203 Q * mire Ping timeout: 480 seconds 1138708256 M * Hollow bonbons: hows your applet skeleton going? 1138708322 M * bonbons not progress this morning, was playing around with my broken laptop-battery (removed dead cells => 3 good out of 8 and hoping I can get the 3 left do their job) 1138708413 M * Hollow heh 1138708439 M * bonbons at least now the laptop agrees to try charging though acpi/battery/*/state says remaining capacity == 0 but encouraging voltage 1138708498 M * bonbons oh, just got a capacity jump: 1269mAh (out of 4800 @ 8 cells => better out of 1800) 1138708708 M * bonbons sh*t, now back at 0mAh (crazy battery) [that laptop as a real plague: broken battery, crazy HDD which only shares data for first 30 minutes] expected cost to get that one working again 250€ 1138708741 M * Hollow you probably better buy a new one ;) 1138708785 M * Hollow ok, so will try to record the mix i planned yesterday, but still wathcing though.. 1138708832 M * bonbons yeah, but need to find a shop where they don't sell things overpriced (they ask 200€ just for battery here) in Germany I saw around 145€ 1138708854 J * lilalinux ~plasma@80.69.35.186 1138708869 M * Hollow well, if you don't find a pleasant importer, i'll do it for you ;) 1138709070 M * bonbons will look again at what is offered where at which prices, but currently I'm hoping AMD wont wait too long for releasing DDR2-Dualcore Turion so I can by it all at once 1138709183 J * FEN_HIN ~JFOC@202.47.70.114 1138709213 Q * FEN_HIN Quit: 1138709389 A * FaUl is satisfied with his Thinkpad X31 1138709761 J * FireEgl Atlantica@2001:5c0:84dc:: 1138709853 N * Lars Guest1888 1138710297 J * mire ~mire@150-167-222-85.COOL.ADSL.VLine.Verat.NET 1138710803 J * phycho ~phych0@209.8.233.208 1138711360 Q * mire Ping timeout: 480 seconds 1138711364 J * kongsted ~ak@0x5551697e.adsl.cybercity.dk 1138711374 Q * kongsted Open Proxy found on your host. Please visit http://www.blitzed.org/proxy?ip=85.81.105.126 for more information. 1138711381 J * mire ~mire@150-167-222-85.COOL.ADSL.VLine.Verat.NET 1138713278 J * thomsen ~thomsen@83-221-78-17.cable2.primacom.net 1138713315 M * thomsen hello, can anybody answer me some questions about vserver ? 1138713733 Q * phycho Read error: Connection reset by peer 1138714685 Q * shedi Ping timeout: 480 seconds 1138714776 Q * thomsen Ping timeout: 480 seconds 1138714922 J * thomsen ~thomsen@83-221-78-17.cable2.primacom.net 1138715042 J * shedi ~siggi@inferno.lhi.is 1138715589 J * phycho ~phych0@209.8.233.208 1138715706 M * thomsen hi , can anybody answer me some questions about vserver ? 1138715733 M * mnemoc thomsen: don't ask to ask, just ask and be patient 1138715889 M * thomsen okay :-) is it possible to run a for ex. a suse-vserver on a debian host ? how can i install it ? ( i only found infos about debian's bootstrap ) 1138715924 M * phycho yeah it is 1138715928 M * phycho dont know how to do it 1138715928 M * phycho but.. 1138715932 M * phycho lycos does it 1138715933 M * phycho =) 1138715937 M * phycho and lycos uses vserver 1138716041 M * thomsen okay ... my friend google dosn't helped me much .. 1138716088 M * phycho yeah googles pretty shit these days 1138716091 M * phycho full of too much crap 1138716279 M * thomsen yes ... 1138716334 J * dearaujo ~dan@cpe-66-25-189-193.austin.res.rr.com 1138716358 Q * dearaujo Quit: 1138716411 J * dearaujo ~dan@cpe-66-25-189-193.austin.res.rr.com 1138716862 Q * balbir Quit: Leaving 1138717064 Q * phycho Ping timeout: 480 seconds 1138717660 J * Doener doener@i5387EBD1.versanet.de 1138718343 M * thomsen does anybody of you how to set up a suse-vserver on a debian host ? 1138718500 M * bonbons thomsen: should be no different from setting up a SuSE-vserver on a suse host! You could even just put a "normal" suse system in the guest and disable mount/network/module prelated parts of boot-scripts 1138718593 M * thomsen is there a how-to some where ? i only found debain/ubuntu/gentoo-stuff 1138718615 J * undefined ~undefined@adsl-68-93-109-94.dsl.rcsntx.swbell.net 1138718711 M * bonbons don't know 1138718815 M * thomsen :-( 1138718947 M * thomsen okay ... i found what i need ( i hope ) - sometimes reading things more carfully helps 1138718952 M * thomsen http://debian.marlow.dk/vserver/guest/ 1138719043 J * lilalinux_ ~plasma@dslb-084-058-211-026.pools.arcor-ip.net 1138719396 M * daniel_hozac suse is not natively supported, so you'll need to go the tarball route. 1138719429 Q * lilalinux Ping timeout: 480 seconds 1138719526 M * thomsen okay .... i see .... i also read about that yast isnt running propper in a vm 1138719605 M * daniel_hozac why not? 1138719737 M * thomsen don't know why .. only read it ... doesn't tryed it yet 1138719991 M * bonbons will certainly only be able to du part of tasks (bootloader, partitionning, ... cannot be done with yast) But once there is a base suse system available yast should be usable for package management 1138720032 M * thomsen okay 1138720048 M * thomsen i think i'll only try it an see what happens :-) 1138720146 M * bonbons that's the best way to find out! 1138720150 M * bonbons good luck 1138720327 M * thomsen thank you ! 1138720519 P * thomsen 1138721606 J * joebullhead ~joebullhe@209.149.59.202 1138721925 Q * harry Read error: Connection reset by peer 1138721977 Q * mire Ping timeout: 480 seconds 1138722222 J * harry ~harry@d515321D1.access.telenet.be 1138723273 J * Viper0482 ~Viper0482@p549756CF.dip.t-dialin.net 1138724738 J * titoo ~titoo@dh3cp39.kt.dtu.dk 1138724852 Q * Viper0482 Remote host closed the connection 1138724870 M * titoo Hello, I have some problems starting a vserver from another host after a rsync 1138724901 M * titoo I am doing incremental backup on the backup system with cp -al and rsync 1138724916 M * titoo from the main host no problem 1138724947 M * titoo from the backup server, no network after the second rsync 1138724957 M * bonbons titoo: what problems do you experience? 1138724978 M * titoo bonbons: the same as last time if you remember 1138724992 M * bonbons I don't remember... 1138724998 M * titoo no network access from/to the vserver and the world 1138725011 M * titoo but ping from the backup host to the vserver ok 1138725026 M * titoo I setup the "prefix" for my interfaces 1138725049 J * Viper0482 ~Viper0482@p549756CF.dip.t-dialin.net 1138725075 M * bonbons what's your intention, just temporarly run a vserver-guest on another host, or completely move it? 1138725095 M * titoo temporarly in case of problem on the main 1138725101 P * undefined 1138725124 M * titoo I have 2 boxes and switch the vserver on the backup before heavy upgrade on the main host 1138725150 M * bonbons so "master" and "slave" host are on same network (different IP bust same prefix/mask) and guest is moved with it's IP address? 1138725196 M * titoo yes 1138725225 M * daniel_hozac titoo: have you set up NAT on the main and not on the backup? or are you in a no-NAT-needed setup? 1138725233 M * titoo http://linux-vserver.org/Recipes You even have my ips at the end of this recipe 1138725251 M * titoo no Nat, I have 3 ips 1138725258 M * titoo fixed routable 1138725259 M * bonbons no overlapping time during which the guest's IP is active on both hosts? 1138725296 M * titoo I do stop on one before start on the other 1138725303 M * titoo should I wait a little ? 1138725318 M * daniel_hozac may be an arp cache issue. 1138725363 M * titoo it would prevent the network to work? 1138725375 M * bonbons waiting a few minutes should be more than enough for ARP caches to get empty. 1138725417 M * titoo ok, I will stop the server and wait and talk to you later 1138725420 M * titoo thanks 1138725420 M * bonbons Did you try IP addresses without the guest? guest's IP on master, ping from outside/to outside; then migrate that IP to the slave and do same test 1138725447 M * titoo I do not understand 1138725469 M * titoo I have one guest 130.225.64.84 1138725477 M * titoo two hosts/masters 1138725480 M * bonbons if this works the problem is not network-related, but maybe incorrectly/incomplete move of guest-config? 1138725485 M * titoo 130.225.64.85 and 130.225.64.86 1138725516 M * bonbons I call the current host "master" and the backuphost "slave", guest just as guest 1138725544 M * titoo my problem is when I migrate 1138725594 M * titoo I try first to wait :) then I will try to understand your test 1138725685 M * bonbons ok [my test is just IP/network level, without vserver-guest; just to get possible causes separated] 1138725791 J * mire ~mire@150-167-222-85.COOL.ADSL.VLine.Verat.NET 1138726012 M * harry anyone here? 1138726024 A * harry has a problem with sync... 1138726035 M * harry and i don't know where the problem is... 1138726043 M * harry http://harry.ulyssis.org/problem.txt 1138726074 M * cehteh whats the prblem? 1138726075 M * harry btw. those are 146GB disks 15k rpm 1138726086 M * harry 2,5MB/s is not enough for 146GB 15k rpm disks 1138726117 M * cehteh syncing is bandwidth shaped 1138726126 M * harry ? 1138726129 M * cehteh there is some interface in /proc/sys to tune it 1138726144 M * harry gandalf:/proc/sys/dev/raid# cat speed_limit_m* 1138726144 M * harry 200000 1138726144 M * harry 1000 1138726156 M * harry 2500 < 200000 1138726172 M * harry and much <<< 1138726201 M * harry sync of md0 and md1 was done at 70MB/s 1138726283 M * cehteh i dont know the exact algorithm which schedules syncing .. increase that limits .. 1138726308 M * cehteh but if ur server is busy with other things, I/O performance might suffer badly 1138726341 M * harry it's doing nothing 1138726345 M * harry nothing at all 1138726358 M * harry gandalf:/proc/sys/dev/raid# cat speed_limit_max 1138726358 M * harry 4000000 1138726359 M * cehteh mhm here it works when i resync 1138726361 M * harry no change 1138726378 M * cehteh be patient 1138726382 M * Doener harry: anything in your syslog? 1138726385 M * cehteh and increase the minimum limit 1138726437 M * cehteh maximum is the upper margin when other I/O is idle, minimum is a minimal rate guranteed even when the server is busy 1138726531 M * harry Doener: no 1138726601 A * harry now created a filesystem on md7 (being a raid0-ed raid1 (over md5/6) 1138726602 M * cehteh patience :) ... better shutdown ur raids properly to avoid resyncing 1138726604 M * harry so a raid10 1138726618 M * harry cehteh: i just created the mirrors 1138726621 M * harry this is no resync 1138726624 M * cehteh ah ok 1138726625 M * harry this is an initial sync :) 1138726637 M * cehteh well you can work while it is syncing .. just no redundancy 1138726674 M * cehteh and afaik the newer kernels have raid10 buildin .. little less overhead and less confusion for volume managers 1138726678 M * titoo bonbons: what is the recommanded way to copy the config of a vserver from master to slave? 1138726711 M * harry cehteh: how do i create a raid10? 1138726719 M * harry docs say: mdadm v 1.7.0 or higher 1138726728 M * harry but my 1.9.0 doesn't seem to support raid10 1138726743 M * bonbons copy /etc/vservers/ from "master" to "slave" and then do eventual changes to reflect differences on both hosts (different network interfaces) 1138726743 M * cehteh mhm .. i cant say 1138726767 M * cehteh maybe ur distributor disabled raid10 support? 1138726780 M * harry cehteh: i'm the distributor :) 1138726803 M * titoo ok simple copy, I will test so. 1138726808 M * cehteh # mdadm --version 1138726808 M * cehteh mdadm - v1.12.0 - 14 June 2005 1138726828 M * cehteh .. and i thought debian is behind := 1138726865 M * bonbons for testing, I'm preparing a second box with vserver-support, will then try to move over one of my guests. If your kernel/utils versions differ to much between both hosts, maybe an update would be recommended 1138727062 Q * gerrit Quit: Client exiting 1138727193 M * titoo bonbons: same hardware, same kernel except an extra network card in the "slave" but not used 1138727356 M * titoo I wonder if the reason could be the multiple hard links due to my backup with cp -al to rotate the backups 1138727391 M * titoo I am doing now without cp -al to test. 1138727401 Q * AllenJB Remote host closed the connection 1138728031 Q * bragon Read error: Connection reset by peer 1138728261 M * titoo nope, does not look like the cp -al was the problem 1138728439 Q * prae Quit: Execute Order 69 ! 1138728461 J * undefined ~undefined@adsl-68-93-109-94.dsl.rcsntx.swbell.net 1138728598 M * titoo bingo ! 1138728610 M * titoo arp problem it looks like 1138728730 M * titoo using 2 shells I started the vserver from the backup doing an arping to update the rooter at the same time 1138728750 M * titoo it worked 1138728799 M * titoo bonbons: I will try to perform the test several times (move from master to slave and return) and keep you informed 1138728832 M * titoo thanks for the time you spent try to help me, I really appreciate this 1138729243 J * stefani ~stefani@superquan.apl.washington.edu 1138729616 M * titoo bonbons: really that, I run "arping -U -I eth1 -c 10 -s 130.225.64.84 130.225.64.1" while booting and it is solving the problem directly 1138729843 Q * Viper0482 Quit: bin raus, 1138729928 J * Viper0482 ~Viper0482@p549756CF.dip.t-dialin.net 1138729973 Q * Viper0482 Remote host closed the connection 1138730044 M * bonbons titoo: fine 1138730093 M * titoo bonbons: 15h in 2 times to understand the problem... but now it works :) 1138730177 J * Viper0482 ~Viper0482@p549756CF.dip.t-dialin.net 1138730202 M * bonbons problems are often simple to solve, just hard to identify the cause or even think about looking at the right place for the cause 1138730777 M * titoo true 1138731870 Q * Viper0482 Quit: bin raus, 1138731955 J * Roey ~katz@h-69-3-4-130.mclnva23.covad.net 1138731956 Q * titoo Quit: Leaving 1138732325 J * Viper0482 ~Viper0482@p549756CF.dip.t-dialin.net 1138732409 Q * shedi Quit: Leaving 1138732754 Q * Hollow Ping timeout: 480 seconds 1138732771 J * Hollow ~hollow@home.xnull.de 1138733309 J * ntrs ~ntrs@68-188-50-87.dhcp.stls.mo.charter.com 1138733309 Q * ntrs_ Read error: Connection reset by peer 1138734719 Q * lilalinux_ Remote host closed the connection 1138734793 J * prae ~benjamin@sherpadown.net 1138734810 J * lilalinux ~plasma@dslb-084-058-211-026.pools.arcor-ip.net 1138735247 J * shedi ~siggi@inferno.lhi.is 1138735341 J * DoM__ ~DoM@151.56.220.161 1138735683 Q * lilalinux Remote host closed the connection 1138735724 Q * DoM_ Ping timeout: 480 seconds 1138735868 J * lilalinux ~plasma@80.69.35.186 1138735987 M * yang Is telnetd disabled from vservers? I can log in with ssh, but telnetd gives me Jan 31 19:31:49 shells inetd[27302]: telnet/tcp: bind: Address already in use 1138736235 M * lilalinux yang: your host is bound to all ip adresses 1138736248 M * lilalinux try to bind it to only one 1138736339 M * yang how 1138736445 M * lilalinux man telnetd 1138736474 M * lilalinux i don't have telnetd installed, it's deprecated 1138736592 M * brc_ yang: disable telnet on the host. 1138736599 M * brc_ and it will work on the guests 1138736622 M * brc_ when the host fires up soem service, it binds it to ALL ips. so when a vserver starts it can not bind the port (host already using) 1138736623 J * Smutje_ ~Smutje@xdsl-87-78-63-149.netcologne.de 1138736637 M * brc_ Solution: disable telnet or make it bind only to the hosts ip (dunno how) 1138736711 M * bonbons you could also put host's services in a host network context (and limit that context to only a few addresses) 1138736749 Q * Smutje Ping timeout: 480 seconds 1138736776 M * Hollow evening folks 1138736941 M * yang well i am using debian package of telnetd 1138736954 M * yang i dont see if there are settings to bind it to certain ip 1138737325 M * yang w 1138737411 M * brc_ i know where to set on another daemons, but not on telnet. 1138737413 Q * joebullhead Quit: joebullhead 1138737417 M * undefined yang: is telnet being ran from inetd? (i can't remember) 1138737733 M * yang no 1138737759 M * yang yes it is - telnet stream tcp nowait root /usr/sbin/tcpd /usr/sbin/in.telnetd 1138738015 M * undefined yang: i don't think inetd allows you to bind to a particular address (i wouldn't mind being proven wrong), but i know xinetd does 1138738054 Q * prae Quit: Pwet 1138738126 M * undefined yang: try "If a string of the form @hostname is appended to the service, it causes inetd to bind to the port for the service on only the specific IP address associated with hostname instead of listening on all available addresses." 1138738137 M * undefined yang: from http://linuxreviews.org/man/inetd/ 1138738164 Q * lilalinux Remote host closed the connection 1138738219 M * Hollow bonbons: any news regarding the applets? 1138738306 M * bonbons not yet, will more probably be for tomorrow evening, late (quite a few things in my pipe) 1138738341 M * Hollow ok, good, i'm already quite tired.. so, we'll see tomorrow :) 1138738357 M * yang ok so it wont work 1138738400 M * bonbons just hope I will get enough progress tomorrow evening, otherwise it's delayed to friday evening or week-end (my two days holliday are over) 1138739037 M * undefined yang: you can switch to xinetd which i know does binding to specific addresses 1138739134 M * undefined yang: i think i remember the debian xinetd package including a perl script to convert an inetd config file to xinetd format 1138739159 J * Aiken ~james@tooax6-013.dialup.optusnet.com.au 1138742245 M * dearaujo I have an guest running that uses ip 192.168.0.55/32 and my host uses 192.168.0.4/24 - can anyone tell me why the /32 is there? 1138742350 M * daniel_hozac because that's the default prefix. 1138742365 M * daniel_hozac change it in /etc/vservers//interfaces/*/prefix. 1138742374 M * dearaujo any reason why I should change it 1138742381 M * dearaujo technically speaking? 1138742425 M * dearaujo i assume since the logical AND of the prefix would and the ip would put me on the actual network? 1138742475 M * bonbons yes, coexistance of two different overlapping networks dosen't work very well (some apps may have bad network connection) 1138742495 M * daniel_hozac only if on different interfaces. 1138742504 M * dearaujo this is one interface 1138742783 Q * Viper0482 Quit: bin raus, 1138742832 J * tudenbart ~willi@xdsl-213-196-248-213.netcologne.de 1138743070 Q * marl Ping timeout: 480 seconds 1138743261 Q * dothebart Ping timeout: 480 seconds 1138744063 Q * mire Ping timeout: 480 seconds 1138744362 J * mire ~mire@150-167-222-85.COOL.ADSL.VLine.Verat.NET 1138744482 Q * bonbons Quit: Leaving 1138744544 Q * shedi Remote host closed the connection 1138744651 J * shedi ~siggi@inferno.lhi.is 1138744728 Q * dearaujo Quit: goodbye 1138745410 J * menomc ~amery@200.75.27.31 1138745574 Q * mnemoc Ping timeout: 480 seconds 1138745574 N * menomc mnemoc 1138747317 N * Bertl_zZ Bertl 1138747322 M * Bertl evening folks! 1138747381 M * blizz hiho 1138747434 M * Bertl hey blizz! 1138747451 M * Bertl Aiken: ping? 1138747651 M * mugwump Bertl: how about that new split! :) 1138747657 M * Bertl ebiederm: hey! your last reply (the slightly sarcastic one :) sounds really familiar :) 1138747678 M * Bertl mugwump: be assured, it will happen today 1138747682 M * mugwump thanks 1138747710 M * Aiken Bertl pong 1138747748 M * Bertl Aiken: ah, good, I installed 2.6.16-rc1-vs2.1.0.8 on my alpha 1138747768 M * Aiken how did it go? 1138747781 M * Bertl the funny thing is, a) it compiled fine (no /dev removal) and it booted without any alignment issue 1138747792 M * Bertl (that was b) 1138747807 M * Bertl soo, now I wonder what is/was the difference ... 1138747810 M * Aiken did the permissions on /dev/null at least change? 1138747820 M * Aiken 3 things happen with /dev/null 1138747822 M * Aiken 1. nothing 1138747825 M * Aiken 2. removal 1138747831 M * Aiken 3. permission change 1138747845 M * Bertl crw-rw-rw- 1 root root 1, 3 2006-01-28 11:59 /dev/null 1138747857 M * Bertl well it was obviously 'touched' but that's about it 1138747862 M * Aiken crw-rw-rw- 1 root root 1, 3 2006-02-01 16:12 /dev/null 1138747883 M * Bertl and I know that this is/was a real issue, as it got fixed 1138747897 Q * Guest1888 Read error: Connection reset by peer 1138747907 M * Bertl so, now I'm confused about the alignment thing ... 1138747949 M * Bertl as far as I could figure, alpha in general should be picky about misalignment (i.e. not just a few chips/versions) 1138747968 M * Aiken rc1 -> the scsi module I want won't compile 1138747985 M * Aiken rc1-git4 -> OOPS but have not managed to capture an oops yet 1138747998 M * Aiken rc1-git4-vs-2.1.0.6.1 alignment issues 1138748024 M * Bertl well, didn't try git4 yet, as the rc1 worked 1138748025 M * Aiken have not tried 2.6.0.8 or so yet as I am waiting for an rc2 or git5 1138748088 M * Bertl what I'd really like to know is which of the alignment attributes was the one fixing the issue 1138748115 M * Bertl because after further thinking about it, I concluded that only the entire struct could be misaligned 1138748147 M * Bertl the pointers inside the struct should be 64bit by default, and the compiler should _automatically_ align them properly 1138748166 M * Bertl (unless the struct was specified as packed) 1138748217 M * Bertl btw, I'm using gcc 3.3.5 and binutils 2.15 1138748257 M * Bertl does that make sense to you? 1138748261 M * Aiken Reading specs from /opt/apps/alphaev5-unknown-linux-gnu/lib/gcc-lib/alphaev5-unknown-linux-gnu/3.3.6/specs 1138748262 M * Aiken Configured with: ../configure --prefix=/opt/apps/alphaev5-unknown-linux-gnu --build=i686-pc-linux-gnu --host=i686-pc-linux-gnu --target=alphaev5-unknown-linux-gnu --enable-shared --enable-threads=posix --enable-languages=c,c++ 1138748262 M * Aiken Thread model: posix 1138748262 M * Aiken gcc version 3.3.6 1138748265 M * harry Bertl: you there? 1138748274 M * harry got a sec? 1138748277 M * Aiken GNU assembler version 2.15.92.0.2 (alphaev5-unknown-linux-gnu) using BFD version 2.15.92.0.2 20040927 1138748279 M * Bertl harry: nope, I'm not here :) 1138748281 M * harry i've got a huge problem 1138748288 M * Bertl harry: sure, what's up? 1138748290 M * harry damn... then i'll check back ;) 1138748292 M * harry ahm... 1138748294 M * harry scsi problem 1138748300 M * harry 1 pci bus 1138748316 M * harry 64bit 133MHz, 2 slots, both containting a lsi21320-r scsi controller 1138748321 M * harry pretty good stuff imho 1138748336 M * harry split backplane with each 2 146GB 15k rpm scsi disks 1138748337 M * Bertl sounds like a dell or so 1138748345 M * harry should be quite nice on performance, right? 1138748347 M * Aiken Bertl http://members.optusnet.com.au/~boddingt/crash-rc1-git4.jpg 1138748358 M * harry raid1 over both controllers 1138748367 M * harry so sda sdb on controller 1, sdc sdd on 2 1138748375 M * harry raid1 over sda and sdc 1138748382 M * harry but... what do we see... 1138748384 M * harry [======>..............] resync = 33.4% (45988224/137516288) finish=615.5min speed=2478K/sec 1138748404 M * harry iostat -kx 2 gives me stuff like: 1138748407 M * harry Device: rrqm/s wrqm/s r/s w/s rsec/s wsec/s rkB/s wkB/s avgrq-sz avgqu-sz await svctm %util 1138748410 M * harry sda 0.00 5.47 0.00 39.80 0.00 4967.16 0.00 2483.58 124.80 12.67 341.50 22.50 89.55 1138748413 M * harry sdb 0.00 5.97 0.00 36.32 0.00 4648.76 0.00 2324.38 128.00 12.63 376.22 26.79 97.31 1138748416 M * harry sdc 6.47 0.00 33.33 0.00 5858.71 0.00 2929.35 0.00 175.76 14.43 454.57 24.66 82.19 1138748419 M * harry sdd 9.95 0.00 30.35 0.00 5476.62 0.00 2738.31 0.00 180.46 12.17 428.39 27.02 81.99 1138748422 M * harry look at the last column... 1138748425 M * harry 90+% utilized 1138748429 M * harry how the hell is that possible? 1138748439 M * harry the machine is doing NOTHING at all 1138748452 M * Bertl well, first, why would a raid controller use software raid? 1138748454 M * harry its a 4way dualcore opteron 2.4GHz 1138748464 M * harry Bertl: not a raid controller 1138748466 M * harry scsi controller 1138748473 M * harry and it has 32 gig ram 1138748482 M * harry so there is no bottleneck afaik 1138748485 P * undefined 1138748489 M * Bertl 64bit machine? 1138748492 M * harry yes 1138748514 M * Bertl well, what throughput do you see on the raw scsi device? 1138748517 M * harry how can a 146GB 15k rpm disk be 90+% utilized when it's doing 2,5 meg/s 1138748532 M * harry see the iostat paste... 1138748558 M * harry http://harry.ulyssis.org/problem.txt 1138748592 M * harry i checked the gandalf:~# cat /proc/sys/dev/raid/speed_limit_m* 1138748592 M * harry 200000 1138748592 M * harry 1000 1138748602 M * harry so it should go faster... 1138748620 M * harry can it be irq? bad cable? something else i'm missing? 1138748627 M * harry there are no errors in dmesg or so 1138748634 M * Bertl could you upload the iostat and maybe a bonnie benchmark for a single scsi target? 1138748648 M * Bertl (the iostat in a properly formatted way) 1138748648 M * harry upload the iostat? 1138748681 M * Bertl the tabbing got messed up, and I don't like to put it right :) 1138748712 M * harry harry.ulyssis.org/iostat 1138748724 M * Bertl Aiken: this is with vanilla? 1138748750 M * harry Bertl: strangest thing... md0 and md1 synced at 60-80 MB/s 1138748765 M * harry then i rebooted, attached network cables, setup networking 1138748766 M * Aiken yes 1138748781 M * harry and then i fixed the other raid volumes... and then this was the situation 1138748788 M * Aiken so I still have to chase problems with vanilla first afertall 1138748804 M * harry btw. the grsec + vserver patches also work on 2.6.14.6 ;) 1138748813 M * harry only failure is the Makefile (du'uh) 1138748869 M * harry i thought: maybe irq problem with network card... but i switched to the onboard network interfaces: didn't help 1138748889 A * harry thinks it has to be hardware... but what 1138748901 M * harry btw. it's a HP DL-585 1138748916 M * Bertl ah, yes, would have been my second guess 1138748922 M * Bertl (hp, not the model) 1138748951 M * harry it's a nice machine... if it would work decently 1138748968 M * harry it looked good at first... 80MB/s in a sync, peak performance to 110MB/s 1138748982 M * harry but now... 2,5MB/s on each disk :S 1138749001 M * Bertl well, looks fishy ... if you want to hire me for tracking down that issue, I'd be interested :) 1138749008 M * derjohn harry, 2.5meg sync speed or hdparm -t ? 1138749009 M * harry hehe 1138749015 M * harry derjohn: sunc speed 1138749017 M * harry sync 1138749019 M * derjohn hp == hardware problem? 1138749024 M * derjohn suck speed ? := 1138749025 M * derjohn ) 1138749028 M * harry derjohn: uhu... 1138749050 M * derjohn 2.5 meg sound like silicon image chip crap 1138749056 M * derjohn .. what chipset? 1138749060 M * harry Bertl: they might want me to change jobs... so there might be a position opening up 1138749063 M * harry ;) 1138749067 M * derjohn -> lspci 1138749068 M * harry derjohn: mpt 1138749080 M * harry 0000:05:0d.0 SCSI storage controller: LSI Logic / Symbios Logic 53c1030 PCI-X Fusion-MPT Dual Ultra320 SCSI (rev 07) 1138749083 M * harry 2 of those 1138749089 M * harry so 4 entries 1138749095 M * derjohn harry, RAID mode or JBOD ? 1138749100 M * harry scsi mode 1138749104 M * harry software raid 1138749121 M * derjohn we are using tons of them in Fujistsu Siemens RX300 stuff ... 1138749132 M * derjohn basically they run fine 1138749135 M * harry uhu 1138749137 M * harry they DID 1138749150 M * derjohn until? 1138749154 M * Bertl Aiken: do you use CONFIG_CC_OPTIMIZE_FOR_SIZE? 1138749157 M * harry like i said: syncing md0 and md1, 80MB/s, peak to 110MB/s 1138749167 M * derjohn harry, what was changed? 1138749172 M * harry derjohn: until i set up networking and started configging remotely 1138749187 M * derjohn btw: RAID sync may drop when there is other stuff to do 1138749189 M * harry i don't know exactly when it went wrong 1138749196 M * harry derjohn: it's doing nothing 1138749205 M * harry http://harry.ulyssis.org/iostat 1138749207 M * harry check it out 1138749208 J * phycho ~phych0@209.8.233.207 1138749212 M * harry check %util there.. 1138749219 M * phycho hey guys 1138749221 M * derjohn hmm hmm ... apt-cache search irqbalance? 1138749226 M * derjohn cat /proc/interrupts? 1138749227 M * Bertl welcome phycho! 1138749241 M * derjohn xeons ? 1138749242 M * phycho how can i set my server up so that it uses swap when the allocated amount of ram is reached? 1138749246 M * phycho say i want to give a user 128meg 1138749248 M * harry LOC: 5063727 5063726 5063725 5063724 5063723 5063722 5063721 5063720 1138749251 M * phycho but they occasionally go over 1138749252 M * harry wtf are those? 1138749258 M * phycho can i let them use a 'fake' swap? 1138749265 M * phycho i dont want processes to be killed. 1138749269 M * derjohn harry, those? 1138749274 M * harry LOC 1138749277 M * Aiken Bertl no 1138749281 M * Bertl phycho: yes, with latest devel version you can set a soft limit 1138749289 M * harry 20: 1048999 0 0 0 0 0 0 70 IO-APIC-level ioc0 1138749293 M * harry 22: 1277482 0 0 0 0 0 0 69 IO-APIC-level ioc2 1138749298 M * Aiken # CONFIG_CC_OPTIMIZE_FOR_SIZE is not set 1138749308 M * phycho ooh 1138749308 M * phycho haha i dont fancy upgrading though 1138749308 M * phycho ill wait till its in -stable =) 1138749309 M * harry those look like the 2 scsi's 1138749316 M * derjohn harry, LOC -> dont know 1138749324 A * phycho is very happy with his vserver atm 1138749330 M * phycho anyway i can get iptables to work within it though? 1138749334 M * Bertl phycho: okay, np, in this case, just raise the limit 1138749334 A * phycho wants firewalling 1138749349 M * derjohn harry, I cannot read .... both IRQ on one CPU? 1138749351 M * Bertl phycho: no, networking happens on the host for now 1138749360 M * phycho strange 1138749369 M * phycho why did iptables work on lycos vserver/ 1138749372 M * Bertl phycho: but, you can do all kind of networking/firewalling there 1138749372 M * phycho *? 1138749375 M * harry come to think of it... 1138749380 M * harry ALL interrupts on 1 cpu 1138749399 M * Bertl phycho: simple, they have a client/server setup, where the iptable commands get sent to a daemon on the host 1138749411 M * phycho ah 1138749422 M * Bertl phycho: that daemon checks the stuff, and adds them if appropriate 1138749435 M * phycho hehe sweet 1138749452 M * phycho ouch at the amount of ram my mail one is taking up 1138749456 M * phycho 49153 36 238.1M 143.9M 0m02s90 0m00s57 15m48s90 mail 1138749457 M * phycho lol 1138749457 M * harry derjohn: from what i know about computers (which is maybe not that much for you guys), it looks like hardware crap 1138749494 M * Bertl phycho: sidenote, you should use static contexts ... 1138749515 M * harry derjohn: irqbalance installing 1138749523 M * phycho static contexts? 1138749530 M * harry but that wouldn't explain why it did work , and it doesn't anymore 1138749550 M * derjohn harry, dont judge so fast :) MPT and symbios are generally fine. where do all the 0 come from in you table? 1138749550 M * Bertl phycho: yes, 49153 is a dynamic context. those will go away soon ... 1138749556 M * derjohn did you use irqbalance? 1138749561 A * harry reboots ;) 1138749565 M * phycho bertl - what exactly is a context? 1138749573 M * Bertl phycho: just add an entry 'context' to your config for each guest, with an unique number 1138749577 M * harry the 0-s??? no interrupts handled by that cpu : 1138749578 M * harry :) 1138749583 M * harry it's a quad dualcore opteron 1138749588 M * Bertl phycho: numbers between 2 and 49151 for now 1138749588 M * harry hence the 8 columns 1138749596 M * derjohn harry, oh, that explains it :) 1138749601 M * phycho ah k 1138749612 M * phycho ill leave it the way it is for now 1138749617 M * phycho if it aint broke.. dont fix it =) 1138749617 M * harry it's gonna be used for virtual servers... it's my end of the year present 1138749618 M * derjohn harry, 8 virtual cpu ... wow ... 1138749634 M * phycho lol sweet 1138749640 M * harry 32 gig ram 1138749642 M * derjohn harry, I want one for free ;) 1138749648 M * phycho ouch 1138749650 M * harry 4x146GB 15k rpm disks 1138749650 M * Bertl phycho: so be it, just remember that note when they are removed in a later version 1138749660 M * phycho hehe will do :) 1138749663 M * harry just for me to play with :) 1138749672 M * derjohn harry, that hedonism! 1138749673 M * phycho probably forget and come in here screaming about why it doesnt work ;) lol 1138749682 M * derjohn harry, which diostro ? 1138749687 M * harry anywya, there is a job opening at our place ;) 1138749690 M * harry derjohn: debian 1138749696 M * derjohn harry, apt-get install irqbalance ? 1138749703 M * harry derjohn: did it allready 1138749711 M * harry now just rebooted for network/... crap 1138749737 M * harry well.. rebooting... lets hope networking works now :s 1138749756 M * harry (it was a bit of a pain in the ass when i last booted it, but i think i fixed it :s) 1138749841 M * derjohn harry, what NIC ? 1138749850 M * harry broadcom tg3 1138749864 M * derjohn harry, i was afraid that it would be such a one. 1138749873 M * derjohn harry, kernel from debian or vanilla ? 1138749874 M * Bertl hehe, always fun :) 1138749881 M * harry 0000:02:06.0 Ethernet controller: Broadcom Corporation NetXtreme BCM5704 Gigabit Ethernet (rev 10) 1138749885 M * harry 0000:07:0a.1 Ethernet controller: Intel Corp. 82546EB Gigabit Ethernet Controller (Copper) (rev 01) 1138749890 M * harry 2 each 1138749899 M * harry derjohn: my kernel :) 1138749900 M * Bertl better get the bcm driver 1138749922 M * derjohn harry, Debian removed the TG3 firmware from the kernel code due to licensing foo. 1138749937 M * harry http://harry.ulyssis.org/patch-2.6.14.4-vs2.1.0-grsec2.1.7_sendfilefix.diff.gz 1138749937 M * phycho http://www.freevps.com/docs/faq.html 1138749938 M * derjohn harry, take bcm manuf driver or vanilla kernel 1138749939 M * harry that patch :) 1138749940 M * phycho this looks nice 1138749941 M * phycho =) 1138749959 M * harry but on a 2.6.14.6 1138749966 M * phycho grsec as well.. 1138749969 M * phycho how did you manage that? 1138749969 M * phycho :P 1138749969 M * harry yup 1138749979 M * phycho i couldnt compile the two patches on mine 1138749979 M * Doener hm, a new openvz release... even against 2.6.15... 1138749979 M * phycho heh 1138749980 M * phycho i tried 1138749982 M * harry phycho: i made it :) 1138749995 M * harry works without any problem : 1138749996 M * harry :) 1138750003 A * phycho saves a copy 1138750009 M * phycho might upgrade my kernel with it later :) 1138750016 M * Bertl Doener: really? 1138750016 M * harry phycho: it stays there... you don't have to backup it :) 1138750036 M * harry phycho: with only a Makefile exception, it applies to 2.6.14.6 too 1138750039 M * Doener http://lwn.net/Articles/169972/ (well, beta) 1138750059 M * Doener i wonder what happened to stable 2.6.8 ;) 1138750082 A * harry hates 2.6.15 kernel 1138750086 M * harry it sucks donkeyballs 1138750086 M * phycho guys 1138750088 M * phycho http://www.freevps.com/docs/faq.html 1138750091 M * harry never seen it work decently 1138750091 M * phycho is that the same as vserver 1138750094 M * phycho except a hack 1138750095 M * phycho ? 1138750103 M * phycho coz a lot of the functions are the exact same 1138750107 M * harry it is 1138750108 M * harry sorta 1138750113 M * phycho thought so.. 1138750116 M * Doener it forked a long time ago 1138750120 A * harry more fond of vserver : 1138750121 M * harry :) 1138750130 M * derjohn Bertl, http://vserver.13thfloor.at/Experimental/patch-2.6.15.1-vs2.1.0.5.1.diff <- is this the last devel for 2.6.15 ? or is it exp ? 1138750149 M * phycho it has features that i want 1138750150 M * phycho :p 1138750150 M * Bertl phycho: well, freevps branched in an early stage ... 1138750151 M * harry btw, derjohn irq balancing didn't solve anything 1138750161 M * derjohn harry, what do those quad machines cost about ? 1138750163 M * phycho i want some of those features in vserver ;) 1138750177 M * Bertl phycho: unfortunately it became 'comercialized' and showed a lack of stability 1138750185 M * phycho ah :( 1138750211 M * Bertl phycho: but if you are going to try it, please let me know how it goes 1138750215 M * phycho k 1138750221 M * phycho i might try it on my other server 1138750224 M * phycho ive got more than one 1138750226 M * phycho :) 1138750275 M * harry derjohn: i think this one was about 23000 euoro 1138750277 M * harry euros 1138750303 M * derjohn harry, your santa had a rellay big bank account ;) 1138750309 M * Bertl derjohn: it is exp, but probably your best choice for 2.6.15 1138750319 M * harry derjohn: end of the year... we had some money leftover at work 1138750325 M * harry we needed a virtualisatiaon machine 1138750343 M * harry so they told me... get yourself a decent machine 1138750347 M * harry so i did 1138750348 M * derjohn harry, np, it you want to get rid of if ... let me be you /dev/moneynull ;) 1138750366 M * harry derjohn: i work at a university 1138750369 M * harry they have money 1138750374 M * harry i don't get paid all that mych 1138750376 M * harry much 1138750380 M * harry but i do get nice toys 1138750389 M * harry like 2 ibm sharks 1138750394 M * harry 2 SAN networks 1138750397 M * derjohn harry, yes, the thing will probably work, but n-way machines tend to be picky ... (in choise of kernels,admins etc) 1138750408 M * harry LTO tape lib, 3494 tapelib, ... 1138750417 M * derjohn harry, real HW Pr0n ! 1138750422 A * phycho needs to find out how to setup vserver disk limit 1138750435 M * harry phycho: lvm? 1138750446 M * Bertl phycho: there are abunch of pages for that on the wiki 1138750448 M * phycho without using lvm 1138750458 M * phycho i dont want to have to rebuild my disks 1138750462 M * harry derjohn: irq balance didn't solve anything :s 1138750470 M * phycho i dont want user limits in the vps per user id 1138750480 M * phycho i just want limits on the host system to prevent a vps using too much space. 1138750480 M * derjohn harry, I read but running out of ideas .... 1138750501 M * derjohn harry, hdparm -t /dev/mdx ? 1138750571 M * harry sec... ;) 1138750586 M * harry it's a slow machine (*ashamed of saying that*) 1138750586 M * Bertl phycho: that's what we call (per guest) disk limits 1138750601 M * derjohn harry, ? 1138750612 M * harry installing hdparm takes some time 1138750616 M * harry Setting up hdparm (6.1-2) ... 1138750619 M * harry allmost there now... 1138750620 M * harry :s 1138750627 M * phycho k 1138750633 M * harry disks are under really heavy load (why, i don't know) 1138750647 M * derjohn harry, btw: apt-get install sysstat 1138750651 M * derjohn harry, iostat -k 1000 1138750657 M * derjohn harry, iostat -k 1 1000 1138750661 M * harry derjohn: you want iostat?? 1138750665 M * Bertl phycho: http://linux-vserver.org/Disk+Limits 1138750674 M * harry http://harry.ulyssis.org/iostat 1138750680 M * harry Timing buffered disk reads: 6 MB in 3.62 seconds = 1.66 MB/sec 1138750682 M * Doener harry: did you touch the scsi cables? i just found some mails on google, where someone bent down a pin in the connector and that caused such a slowdown for him 1138750687 M * derjohn harry, sure, look how many tps hit the disk in the backgrund 1138750688 M * Bertl phycho: btw, you'll need static context ids for that to work on a shared partition :) 1138750696 M * phycho k ty. 1138750703 M * phycho k 1138750711 M * Doener (yes, i know this sounds like an incredibly stupid suggestion ;) 1138750711 A * harry will open the machine tomorrow at work 1138750723 M * harry i'll check there... looks like there is some HW problem 1138750731 A * harry will check/change the cables etx... 1138750732 M * derjohn harry, the value is definitly slow .... 1138750739 M * harry derjohn: no kidding :) 1138750749 M * derjohn harry, hdd in frames? 1138750757 M * harry as i said earlier... md0 and md1 synced at 60-80 MB/s 1138750757 M * derjohn harry, backplane? 1138750785 M * harry i'm at home now, and the machine is at work 1138750788 M * harry can't do hardware checks 1138750795 M * harry but will do them tomorrow 1138750878 P * meandtheshell 1138750949 M * harry time to go to bed... 1138750955 M * derjohn harry, n8 1138750959 M * harry things are getting hilarious heer... 1138750960 M * harry here 1138750968 M * harry http://harry.ulyssis.org/hdparm 1138750974 M * harry read... and cry :s 1138750989 M * phycho lol 1138751036 M * harry cry, not laugh ;) 1138751036 M * Doener ouch! 1138751064 M * harry reload 1138751072 M * harry did it with time now... 1138751427 P * stefani I'm Parting (the water)