1154305626 J * juggo ~lemur@h-68-166-181-4.sttnwaho.covad.net 1154305789 Q * Pazzo Quit: Ex-Chat 1154306446 J * shedi ~siggi@inferno.lhi.is 1154307050 Q * x03 Ping timeout: 480 seconds 1154309131 Q * kir Ping timeout: 480 seconds 1154310153 M * Skram new util-vservers? 1154310165 M * daniel_hozac really? 1154310175 M * daniel_hozac where? 1154310336 M * Skram i saw it in portage 1154310345 M * Skram heh 1154310346 M * Skram i tink 1154310347 M * Skram one sec 1154310362 M * Skram [ebuild U ] sys-cluster/util-vserver-0.30.210-r17 [0.30.210-r15] 26 kB 1154310381 M * daniel_hozac so not really then. 1154310387 M * Skram oh 1154310657 J * gerrit ~gerrit@66.46.95.226 1154310986 Q * FireEgl Quit: Bye... 1154311370 J * FireEgl ~FireEgl@Atlantica.Tcldrop.Com 1154311656 Q * Radiance Remote host closed the connection 1154312775 J * Radiance e12f963992@halt.1984world.eu 1154314892 J * meandtheshell ~markus@85-124-36-115.dynamic.xdsl-line.inode.at 1154315729 J * KantankerousKid ~blah@tor-irc.dnsbl.oftc.net 1154318565 Q * KantankerousKid 1154319796 Q * micah Ping timeout: 480 seconds 1154322491 J * yarihm ~yarihm@whitehead2.nine.ch 1154323360 M * cehteh Bertl_oO: ping 1154323488 J * micah elijah@micah.riseup.net 1154324632 J * zkbrsnie ~zkbrsnie@83-64-146-226.klosterneuburg.xdsl-line.inode.at 1154325951 J * Zaki[] ~Zaki@212.118.105.126 1154326261 J * coocoon ~coocoon@p54A070EA.dip.t-dialin.net 1154326267 M * coocoon morning 1154326338 Q * Zaki_ Ping timeout: 480 seconds 1154326774 J * dna ~naucki@dialer-161-104.kielnet.net 1154328525 Q * lilo2 Remote host closed the connection 1154328587 J * lilo2 ~0710AAD4@tor-irc.dnsbl.oftc.net 1154329113 J * Viper0482 ~Viper0482@p54975F93.dip.t-dialin.net 1154329138 J * renihs ~penguin@83-65-34-34.arsenal.xdsl-line.inode.at 1154329436 Q * complexmind Quit: using sirc version 2.211+KSIRC/1.3.12 1154329451 Q * Aiken Ping timeout: 480 seconds 1154329577 Q * lilo2 Quit: brb 1154329635 J * lilo2 hiddenserv@tor.noreply.org 1154330317 Q * MrX Quit: urk IRC v0.-1.4 - http://urk.sf.net/ 1154332919 J * Brainkill ~NNSCRIPT@tor-irc.dnsbl.oftc.net 1154333053 J * bonbons ~bonbons@83.222.36.236 1154334966 J * pisco ~pampel@p5087BFB3.dip0.t-ipconnect.de 1154335277 J * chand ~chand@gw.net81-65-27.noos.fr 1154335431 M * blues re 1154335465 M * blues Bertl_oO: it seems that problem similar to mine was reported earlier on vserver list :) 1154335488 M * blues Bertl_oO: http://www.paul.sladen.org/vserver/archives/200410/0020.html 1154335764 J * Milf ~Miranda@ipsio83.ipsi.fraunhofer.de 1154335802 J * schimmi ~sts@host10.natpool.mwn.de 1154336121 Q * meandtheshell Quit: bye bye ... 1154336619 M * h01ger moin 1154337115 Q * bonbons Ping timeout: 480 seconds 1154337285 J * kir ~kir@swsoft-mipt-nat.sw.ru 1154337320 J * lilalinux ~plasma@dslb-084-058-197-134.pools.arcor-ip.net 1154337331 M * h01ger if i disklimit a vserver with /e/v/fuu/dlimits, the number of used blocks isnt correct and isnt updated 1154337339 M * h01ger what am i missing? 1154337345 J * bonbons ~bonbons@83.222.36.236 1154337348 M * h01ger (stable vserver..) 1154338110 N * schimmi derSchimmi 1154338150 N * derSchimmi derHerrSchimanski 1154338375 P * derHerrSchimanski Verlassend 1154340242 M * doener h01ger: did you mount the vserver partition with tagxid? 1154340426 M * h01ger yes 1154340708 J * lilalinux_ ~plasma@h1-gw.of.net-lab.net 1154340724 Q * coocoon Quit: KVIrc 3.2.0 'Realia' 1154340830 J * coocoon ~coocoon@p54A070EA.dip.t-dialin.net 1154341071 Q * lilalinux Ping timeout: 480 seconds 1154342021 Q * yarihm Quit: Leaving 1154342133 Q * juggo Quit: Download Gaim: http://gaim.sourceforge.net/ 1154342529 M * renihs hmm whats the max number of vservers running? 1154342677 M * kir renihs, that depends on load 1154342690 M * kir renihs, i.e. on what you want to run in each vserver 1154342702 M * renihs no i mean, how much are possible 1154342715 M * kir renihs, in lab you can run hundreds of those, or even thousands...the main limit is RAM 1154342730 M * renihs there is no hard limit in vserver contexts? 1154342732 M * kir renihs, in the real world situation you probably end up with tens of those 1154342747 M * renihs i have 48gb ram and a few cores availiable 1154342754 M * Zoiah The XID is a 16-bit number, so I say no more than 65535 ;) 1154342758 M * renihs :) 1154342760 M * renihs heh, k 1154342764 M * kir renihs, I'm not sure about vserver, but in openvz we were able to run like 4000 or so (do not remember the actual figure at the moment) 1154342774 M * waldi hmm, can I assign 127.0.0.1 to a vserver? 1154342816 M * waldi Zoiah: and the upper 16k are reserved for dynamic contexes 1154342833 M * kir renihs, let me think...on 48gb ram it would be like 7500 simple vservers (I'm speaking of OpenVZ, Linux-VServer should be very similar) 1154342855 M * Zoiah You can run a vserver as bare as you want. If you just want it to run sh, you can, so I'm sure you can do more. ;) 1154342938 M * kir renihs, here by "simple vserver" I mean the one with init, sshd, crond, syslogd, sendmail, and apache -- under some load 1154343009 M * renihs i get your drift :) 1154344677 Q * coocoon Ping timeout: 480 seconds 1154345265 J * coocoon ~coocoon@p54A0536C.dip.t-dialin.net 1154345482 J * Pazzo ~thomas@dialin-225136.rol.raiffeisen.net 1154346086 Q * Brainkill Server closed connection 1154346101 J * Brainkill ~NNSCRIPT@tor-irc.dnsbl.oftc.net 1154346123 Q * Brainkill Server closed connection 1154346271 Q * shedi Quit: Leaving 1154346386 J * Brainkill ~NNSCRIPT@tor-irc.dnsbl.oftc.net 1154346388 Q * Brainkill Server closed connection 1154346530 Q * kir Quit: Leaving 1154347035 J * mire ~mire@34-166-222-85.COOL.ADSL.VLine.Verat.NET 1154349296 Q * mire Remote host closed the connection 1154350386 Q * gerrit Ping timeout: 480 seconds 1154351075 M * derjohn when were Linux Capabilities introduced? This was already in 2.4 times,but VS supports them only since 2.6 ? 1154351103 M * sid3windr don't think so 1154351114 M * sid3windr vs1.2 also supported it, no? 1154351125 M * derjohn sid3windr, I never used it ;) 1154351130 M * sid3windr hmm 1154351139 M * derjohn but they were introduced with 2.4? 1154351147 M * sid3windr I have no idea :) 1154351188 M * derjohn sid3windr, we share something common ;) 1154351210 M * sid3windr \o/ 1154351495 Q * wenchien Quit: Terminated with extreme prejudice - dircproxy 1.0.5 1154351896 J * jesse_ ~wenchien@221-169-69-23.adsl.static.seed.net.tw 1154351919 N * jesse_ wenchien 1154352078 J * shedi ~siggi@dsl-og-108-50.du.vortex.is 1154352086 M * shedi hi channel 1154352116 M * shedi how would I go about running a ntp-server inside a vserver 1154352128 M * shedi is it possible 1154352143 M * sid3windr well, you'd need to give it access to manipulate the clock on your server 1154352222 M * shedi I have put in the following CAPs CAP_SYS_TIME, CAP_SYS_RAWIO, CAP_NET_RAW, CAP_SYS_RESOURCE 1154352332 M * shedi is the problem related to the ntp server listening in on interface wildcard, 0.0.0.0#123 1154352354 M * shedi because the option to bind it to an address is unavailable 1154352389 M * shedi or am I missing a CAP 1154352485 J * pisc1 ~pampel@p50878A74.dip0.t-ipconnect.de 1154352527 M * sid3windr hmm 1154352535 M * sid3windr listening on the wild card shouldn't be a problem 1154352543 M * sid3windr vserver automatically translates it to the ip address of the vserver 1154352548 M * sid3windr but then I don't know what could be wrong 1154352553 M * sid3windr do you have an ntpd running on the host too? 1154352562 M * shedi no 1154352730 Q * pisco Ping timeout: 480 seconds 1154352968 M * daniel_hozac h01ger: are the files tagged with the xid? 1154353329 M * h01ger no 1154353338 A * h01ger remembers something vaguely 1154353369 M * h01ger i tag them once (how?) and all new files are tagged automatically after? 1154353511 M * derjohn h01ger, chattr ? 1154353517 M * daniel_hozac chxid. 1154353527 A * derjohn slaps himself 1154353590 M * daniel_hozac http://linux-vserver.org/Disk+Limits describes disk limits pretty well. 1154353656 Q * anonc Ping timeout: 480 seconds 1154353686 M * h01ger daniel_hozac, right (that page looks good indeed), thanx! 1154353695 M * h01ger (you pointed me to the flower page :) 1154353755 M * daniel_hozac yep, i figured micah had everything but the configuration already. 1154353805 M * derjohn doener, ping ... 1154353822 M * doener pong 1154353846 M * derjohn daniel_hozac, do you know the current statius about performance benchmarks? What was about "less then 1% overhead" ? 1154353857 M * derjohn doener, fun in proofreading 2 pages, in German ? 1154353886 M * derjohn doener, I fear that I got too technical for that artcile, but .... 1154353921 N * Bertl_oO Bertl 1154353924 M * Bertl morning folks! 1154353939 J * dna_ ~naucki@dialer-161-104.kielnet.net 1154353942 M * h01ger moin moin Bertl 1154353947 M * coocoon Bertl: morning bertl 1154353956 M * derjohn Bertl, EHLO! 1154353957 M * bonbons morning Bertl! 1154353958 M * phedny hi Bertl 1154353959 M * Bertl derjohn: we have no measurable overhead atm 1154353971 M * doener derjohn: sure 1154353972 M * derjohn Bertl, proofs? Links? 1154354003 M * Bertl kernel folks did a lot of testing ... but probably I should say 'no measured overhead atm' :) 1154354034 M * Bertl because I'm pretty certain with the proper tests (not realistic ones though), you can measure some overhead somewhere 1154354050 M * derjohn Bertl, that means ONE guest alone on a host compared to the same system without being a vserver? 1154354084 M * Bertl yep, precisely, on a multi guest system, we can achieve more than 100% native speed :) 1154354115 M * derjohn Bertl, that is true for the devel branch? or for stable too? I remeber you saying the devel scheduler is faster ... I could be completely wrong though. 1154354120 M * bonbons Bertl: except for network when there are plenty IP addresses per guest <- linear/quadratic overhead 1154354132 M * Bertl derjohn: devel, stable shows something around 1% 1154354152 M * Bertl (i.e. the next 'stable' one has no measured overhead :) 1154354153 M * derjohn bonbons, is that _measrured_ ? /me has systems with 64 IP per guest ;( 1154354178 M * bonbons derjohn: not measured, but just expected from code 1154354194 M * derjohn bonbons, pure theory ;) 1154354199 M * derjohn 1154354213 M * Bertl yes, maybe it is not even measurable for 16ips (the current default limit) 1154354365 Q * dna Ping timeout: 480 seconds 1154354387 M * Bertl derjohn: regarding article: you had all the work, so IMHO you should get all the fame too ... unless you still need some help there ... 1154354403 J * wam ~wam@proxy2.msh.de 1154354410 M * bonbons Bertl: did you already play with seq_file for the proc area? 1154354414 M * Bertl welcome wam! 1154354428 M * derjohn Bertl, Doener: well I just need some qualified Feedback. I'll send it as PDF to both of you. 1154354430 M * Bertl bonbons: nope, still on my todo list .. want to attack that? 1154354432 J * schimmi ~sts@port-212-202-73-176.dynamic.qsc.de 1154354443 M * Bertl derjohn: k, np 1154354464 M * bonbons I looked in that corner during my hacking on network code 1154354513 M * bonbons writing to them and the like I get an idea how it works, but mapping them to the proc inodes not yet 1154354591 M * wam Hi, I have some strange behavour: I use openvpn to connect two hosts (sa1, sb1). From the hosts, the vpn is running and all connects work. From a guest on the openvpn-client, I can connect to the guests on the openvpn-server. But I can't connect from the openvpn-server-guests to the openvpn-client-guests. Except I give them an interface with IP=10.111.0.1, dev=tun0, nodev, prefix=24. Then everything works great. But I don't want to give this interface to every g 1154354591 M * wam uest. Any ideas what could be the reason? 1154354616 M * daniel_hozac bonbons: fs/proc/base.c:mounstats_open, fs/namespace.c:show_vfsstat for an example :) 1154354633 M * Bertl okay, I'm still not getting any vserver mails .. could somebody who is subscribed and still gets them bounce them to me? please speak up so that I do not receive them dozents of times 1154354660 M * daniel_hozac Bertl: i had to reactivate my subscription. 1154354677 M * derjohn wam, you need tunX per guest. you an set the tun before starting the guest. 1154354689 M * daniel_hozac http://list.linux-vserver.org/mailman/listinfo/vserver logged in, and it said delivery disabled at the top. 1154354699 M * Bertl ah, cool, let's try that ... (missing mails since 29th) 1154354706 M * coocoon Bertl: this one Vserver Digest, Vol 28, Issue 57? 1154354735 M * Bertl coocoon: I try to answer to them, so a pure digest is not what I like 1154354738 M * wam derjohn: so i must guarantee, that tun is running before any guests? 1154354751 M * derjohn wam, on my FAQ on the wiki there is a desription of the setup. 1154354763 M * coocoon can send all from 29th to u if u want 1154354769 M * wam derjohn: aaah - i thought that was for running openvpn inside the guests... 1154354770 M * derjohn wam: some_hints_from_john, one of the last Qs 1154354779 M * derjohn wam, yes... 1154354780 M * wam thanks ;) 1154354793 M * derjohn mom, maybe i misuderstood your problem ? 1154354825 M * Bertl daniel_hozac: how did you log in there? 1154354827 M * wam derjohn: my openvpn runs on the host. 1154354837 M * derjohn wam, the server? 1154354856 M * daniel_hozac Bertl: field at the bottom, enter email address (Unsubscribe or edit options). 1154354881 M * wam derjohn: I have two hosts - one for the server, one for the client. Both openvpn-processes don't run on VServers, but in context 0. 1154354883 M * daniel_hozac Bertl: then it asked for the password, IIRC. 1154354888 M * Bertl ah, ic ... tx 1154354913 M * wam derjohn: an now I want that the vservers (which run on every host) can access each other via the vpn. 1154354915 M * derjohn wam, ok, and that openvpns run? how does it correlte to te guest? 1154354927 M * derjohn race condition ... 1154354956 M * derjohn hm, well, you have to set routes through the vpn tunnel 1154354971 M * wam derjohn: the routes should be ok I think... 1154355008 M * wam derjohn: I need names here ;) Host A runs openvpn server, Host B runs openvpn client 1154355017 M * wam Both run vservers. 1154355027 M * derjohn if you have 192.168.0.1/24 on host A 192.168.1.1/24 on B, you can try to ping -i 192.168.1.1 from B to A and check qith tcpdump if it gets through 1154355044 M * wam derjohn: I can connect from vservers on Host B to vservers on Host A, but not the other way (normally). 1154355068 M * derjohn sure the connection is via the VPN? 1154355069 M * wam derjohn: but it works, as soon as I create an interface tun0/10.111.0.1/24/nodev on a vserver on Host A. 1154355085 M * derjohn wam why? 1154355094 M * wam derjohn: I even don't have to restart the vserver - vserver ... enter is enough. 1154355102 M * daniel_hozac wam: NAT? 1154355107 M * derjohn simply give the vservers an IP on eth0. 1154355109 M * wam daniel_hozac: no naat. 1154355119 M * wam derjohn: they have one. 1154355123 M * derjohn the routing process will push their requests through the tunnel 1154355137 M * derjohn no need for tunX in this setup. 1154355141 M * doener Bertl: did coocoon bounce the mails? if not I can upload a mbox as well 1154355144 M * wam derjohn: so theoretically the vserver doesn't have to know the route? Will the host route the packets? 1154355173 M * derjohn wam, yes, type an "ip addr" on the host: you'll see all IPs from all guests there. 1154355185 M * derjohn the host (kernel) does the routing. 1154355190 M * Bertl doener: it probably takes some while (greylisting) until I will receive anything ... give me a few minutes :) 1154355236 M * wam derjohn: Here's the experiment: Vserver "i" on Host A: # ping : TIMEOUTS; 1154355271 M * derjohn on the host B try "ip route get 10.111.0.1" 1154355275 M * wam derjohn: when I leave the vserver, add a new interface "dev=tun0; ip=10.111.0.1; nodev; prefix=24" to the vserver on Host A and re-enter it, this ping works. 1154355284 M * derjohn wam, where does it have its gateway? 1154355284 M * wam derjohn: mom 1154355288 M * Bertl doener: nah, doesn't seem so, but a bounce of those 6 emails would be nice ... 1154355334 M * derjohn wam, better to restart the guest ... or at least the daemons within it. 1154355335 M * daniel_hozac wam: are you using old utils or dynamic contexts? 1154355340 M * coocoon Bertl: ah hehe i thouht u won't have it 1154355342 M * coocoon sorry 1154355351 M * Bertl coocoon: np 1154355355 M * doener Bertl: http://www.13thfloor.at/~doener/vserver.mbox 1154355361 M * Bertl okay, tx 1154355363 M * doener contains mails from 28th to now 1154355389 M * wam daniel_hozac: 0.30.209 1154355397 M * wam daniel_hozac: static contextx. 1154355405 M * daniel_hozac so old utils then :) 1154355504 M * wam derjohn: # ip route get 10.111.0.1; 10.111.0.1 via 10.111.0.13 dev tun0 src 10.111.0.14 1154355530 M * derjohn and vice versa? 1154355550 M * wam # ip route get 10.111.0.1; local 10.111.0.1 dev lo src 10.111.0.1 1154355565 M * wam last one was from Host A 1154355599 M * blues Bertl: hello :) 1154355617 M * blues Bertl: what time is there where you live? :D 1154355624 M * doener BUT 1154355640 M * Bertl 16:20 CEST or morning BUT :) 1154355641 M * doener there's no exact time in that timezone ;) 1154356357 M * harry root 11462 0.0 0.0 3920 1832 pts/0 D 16:24 0:00 mkfs.reiserfs -q /dev/vservervg/webmail5lv 1154356361 M * harry damn 1154356569 J * meandtheshell ~markus@85-124-37-20.dynamic.xdsl-line.inode.at 1154356712 M * Bertl back shortly ... 1154357073 Q * Pazzo Quit: Ex-Chat 1154357082 M * daniel_hozac anyone know how far away 2.6.17.8 is? 1154357093 Q * shedi Quit: Leaving 1154357127 M * Milf as an IP-Address or what? 1154357152 M * daniel_hozac more like ETA until it appears on kernel.org ;) 1154357157 M * Wonka lol 1154357172 M * Milf Oh, you mean a kernel revision :) 1154357195 J * MrX ~urk@219.95.238.72 1154357249 M * Milf Sorry, haven't slept with any kernel developer who could have told me that lately :) 1154357256 M * Wonka PING 2.16.7.8 (2.16.7.8) 56(84) bytes of data. 1154357257 M * Wonka From 188.1.144.70 icmp_seq=1 Time to live exceeded 1154357282 M * daniel_hozac lol 1154357283 M * doener wow, kernel 188... that's way into the future :) 1154357374 J * ekc ~nospam@netblock-66-245-252-180.dslextreme.com 1154357376 M * daniel_hozac hopefully :) 1154357719 M * doener daniel_hozac: at least no review cycle for .8 yet. so it's either a few weeks or whenever a new exploit comes aroung the corner 1154357762 M * ekc Is there a way to stop a vserver by using only the ctx id? I accidentally deleted /etc/vservers and vserver-stat is reporting: "11 3 4.8M 1.3M 0m00s00 0m00s00 1h07m38" 1154357795 M * ekc So, I have the ctx # but not the vserver name. 1154357849 M * Zoiah chcontext --ctx 11 killall5 -9 1154357850 M * Zoiah Maybe? :) 1154357874 M * doener vkill? 1154357948 M * ekc yes! vkill works. vkill --xid 11 1154357950 M * ekc thanks 1154357955 M * wam derjohn: solved / but not great news. The vservers use the wrong interface for communication over the hosts vpn tunnel, as they don't know about tun0. So they try to send via eth0, where the default route is. But those packets will be doomed. When I explicitly define the sender's address, all is good. e.g.: ping -I 10.112.0.32 10.111.0.14 works. ping 10.111.0.14 doesn't :-/ 1154358040 M * derjohn wam, vservers only have eth0, nor? 1154358052 M * wam derjohn: but the strange thing is: When I run ip route get on the vserver (which needs pings -I switch), I get # ip route get 10.111.0.14; 10.111.0.14 via 10.111.0.2 dev if7 src 10.111.0.1 1154358058 M * wam derjohn: they have eth0 and eth1 1154358081 M * derjohn if7 ? 1154358095 M * doener name is unknown to the vserver 1154358105 M * wam derjohn: yeah - if7 seems to be tun0. When I run "route -n", the tun0-routes have a asterisk as device-name 1154358106 M * derjohn doener, looks like .. 1154358133 M * wam derjohn: that's why all works when I add the tun0-device via the interfaces-config to the vserver. 1154358139 M * derjohn wam, what IPs are on the tunnel? Did you use a /30 net ? 1154358152 M * wam derjohn: the tunnel is 10.111.0.0/24 1154358157 M * wam ah 1154358160 M * wam damn - sec 1154358161 M * derjohn wam, thats rubbish 1154358165 M * derjohn hm ? 1154358165 M * wam yeah 1154358187 M * derjohn make a /30 on the tunnel, e.g. 10.10.10.1/30 on A and 10.10.10.2/30 on B 1154358192 M * wam derjohn: inet addr:10.111.0.14 P-t-P:10.111.0.13 1154358210 M * wam so it's 10.111.0.12/30 1154358221 M * derjohn then 'ip ro add 10.111.0.0/24 via 10.10.10.2' ... 1154358236 M * derjohn and the vice versa on the other host. 1154358244 M * derjohn bind the 10.111.0.0/24 on the0 1154358246 M * derjohn *eth0 1154358274 M * derjohn to the vserver get 'their' IP addr on eth0, and the host routes _through_ the VPN tunnel 1154358304 A * wam sinks into deep thought 1154358323 M * doener derjohn: in our good tradition of saving keystrokes make that "ip r a ..." :) 1154358335 Q * Milf Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org 1154358338 M * doener SCNR 1154358368 M * derjohn doener, sure, but I didnt know if wam understands it ;) 1154358383 M * derjohn .. one readon more to "ip r a" :) 1154358440 M * doener *lol* took me almost ten seconds to realize that that needs s/d/s/ ;) 1154358456 M * derjohn *reason 1154358463 M * derjohn :) 1154358522 M * derjohn doener, got the PDF ? 1154358532 M * derjohn (i mean: the first one ;)) 1154358987 Q * wam Quit: Verlassend 1154359045 M * doener yep 1154359097 J * kir ~kir@swsoft-mipt-nat.sw.ru 1154359160 M * harry how can i kill a dead process? ;) 1154359164 M * harry (weird q, huh ;)) 1154359191 M * harry i know it's allready dead , but it still has some resources...which i need! 1154359214 M * kir harry, steal those from that process, using steal(2) syscall ;) 1154359239 M * harry hehe 1154359346 M * harry i don't steal from the dead! 1154359351 M * harry they have to parish themselves! 1154359378 M * harry off course pattaching doesn't seem useful either... 1154359586 P * meandtheshell 1154359972 M * harry http://pastebin.ca/108543 1154359998 M * harry i really don't get it 1154360019 M * harry it works perfectly on a similar system... raid10 over 4 146GB scsi disks with lvm2 on it 1154360022 M * harry same kernel 1154360033 M * harry same version of all software 1154360036 M * harry i hate this! 1154360038 Q * waldi Quit: maintenance 1154360063 M * derjohn well, sauron plays with it ;) 1154360071 M * harry hehe :) 1154360084 M * harry that name is chosen for a reason 1154360085 M * harry :) 1154360144 M * Bertl what's the problem? 1154360153 M * Bertl it just asks for a key, no? 1154360385 J * s0undt3ch ~s0undt3ch@bl7-252-144.dsl.telepac.pt 1154360388 Q * s0undt3ch 1154360395 Q * MrX Quit: urk IRC v0.-1.4 - http://urk.sf.net/ 1154360426 J * meandtheshell ~markus@85-124-37-20.dynamic.xdsl-line.inode.at 1154360454 J * s0undt3ch ekqhpe@bl7-252-144.dsl.telepac.pt 1154360477 J * stefani ~stefani@tsipoor.banerian.org 1154360559 M * doener OT: how to turn off greedyness of .* with sed? *can never remember that* 1154360637 M * derjohn doener, you mean how to turn on? 1154360656 M * doener hm, no, off... it seems to be greedy by default 1154360661 M * derjohn doener, i would gues .** is greedy 1154360665 M * derjohn *guess 1154360714 M * doener nvm, i could expand the pattern slightly to workaround it 1154360721 M * doener .** doesn't change anything 1154360823 M * Bertl if you find out, let me know, in (perl)regexp it's /u 1154360907 M * daniel_hozac i thought perl used .*? 1154360935 M * Bertl that means 'optional' /u means ungreedy 1154360961 M * derjohn http://www.regular-expressions.info/repeat.html -> Laziness Instead of Greediness -> ? is to ungreedy it. 1154361006 M * doener daniel_hozac: hm, vim does use that syntax for ungreediness IIRC 1154361123 M * Bertl okay, so let's take an example here: we have ^[abc]+[bc]+$ and want it to match (the first one) ungreedy against aabcbccc 1154361137 M * Hollow wtf! 1154361138 M * daniel_hozac .* is already optional. 1154361162 M * daniel_hozac .*? ungreedies it here. 1154361162 M * Hollow my application to university got rejected 1154361163 M * Hollow :( 1154361174 M * daniel_hozac how come?! 1154361197 M * doener Hollow: NC too bad? 1154361257 M * Hollow my grades were too bad, but the annoying thing is that i had to write a motivation essay, so that with bad grades you can still get it... just to get to know in the end of my inerview that my grades are too bad 1154361258 M * Hollow wtf 1154361270 M * Hollow they could have told me weeks ago that my grades are too bad and not waste my time 1154361271 M * doener Bertl: should be ^[abc]+?[bc]+$ in vim (and probably perl) 1154361308 M * bonbons Bertl: include/net/route.h:150 ip_find_src() What's the exact aim of this, how should be transpose this for a guest which has address range a.b.c.d-e.f.g.h? 1154361310 M * Hollow well, i applied in berlin as well, they don't have such crap 1154361315 M * Hollow and they take nearly everybody 1154361318 M * Hollow :P 1154361341 M * daniel_hozac Hollow: lol, that sucks. 1154361368 M * Bertl doener: and that works? i.e. the first regular expression matches 'aa' only? 1154361380 M * Bertl (the first subexpression that is) 1154361445 M * doener Bertl: should match 'a' only, ungreedy so it stops when that subexpression is fulfilled (+ means 1 or more, right?) 1154361451 M * Hollow daniel_hozac: i even had to drive to uni at 1pm to get a interview at 5pm, and i had to bring the essay personally some weeks ago, so they wasted about 30EUR of my gasoline, and hours of hours of waiting, to reject it based on things they knew ages before 1154361472 M * Bertl doener: a is not matched by the second subexpression, so aa is the smallest match 1154361474 M * daniel_hozac ouch. 1154361482 M * doener ah, right 1154361500 M * Hollow indeed... fcuk the system ;) 1154361548 M * daniel_hozac hehehe. 1154361593 M * daniel_hozac bonbons: select the appropriate address from the available set? 1154361617 M * daniel_hozac (and to make sure it doesn't try to use addresses not available to the guest) 1154361866 N * Ben_zZZz Belu 1154361899 M * bonbons I don't really get the first part in there... if (fl->fl4_src == 0) { } 1154361930 Q * ekc Ping timeout: 480 seconds 1154362016 M * derjohn Hollow, and now? what is plan B ? 1154362038 M * daniel_hozac bonbons: i.e. an unbound socket. 1154362051 M * daniel_hozac bonbons: so it selects the most appropriate one (considering routes and such). 1154362055 M * daniel_hozac (IIUC) 1154362191 M * doener Bertl: ok, I either mis-recalled it or it changed from vim 6.4 to 7, this one works: s/^\([abc]\{-1,}\)[bc]\+$/\1/ 1154362200 M * Hollow derjohn: well, my two applications in berlin are still pending 1154362235 M * doener Bertl: {1,} is like +, {-1,} is like + but matches as few as possible 1154362341 M * Hollow Bertl: regarding the new wiki.. do you mind if i announce it on the ML and ask for help migrating the content? 1154362394 Q * meandtheshell Quit: bye bye ... 1154362439 M * derjohn new wiki? /me saves his FAQ page ... 1154362540 M * Hollow derjohn: http://vwiki.croup.de 1154362590 M * harry Bertl: ? 1154362597 M * harry < Bertl> it just asks for a key, no? 1154362638 M * derjohn and you are the croupier? :) 1154362645 M * Hollow lol.. 1154362655 M * Hollow this is actually the abbreviation of creative group 1154362666 M * kir Hollow, new wiki looks nice 1154362699 M * kir Hollow, s/Older Anouncements/Older Announcements/ on a main page 1154362704 M * Hollow kir: it was a hard time adapting monobook ;) 1154362725 M * Hollow really crappy template.. 1154362728 M * kir Hollow, I see. What will you do when time will come to upgrade the mediawiki? 1154362747 M * Hollow well, i just changed the css 1154362756 M * kir Hollow, well, that might work :) 1154362758 M * Hollow and the location of the search box in monobook.php 1154362777 M * derjohn Hollow, how to access the old wiki? 1154362785 M * Hollow the css should work well, the box has to be moved everytime we upgrade i guess 1154362818 M * Hollow derjohn: well, first the new wiki has to be filled, but then we keep it at like old.l-v.org or so 1154362826 M * Hollow that has to be decided later on 1154362858 M * derjohn k 1154362866 M * Hollow but i seriously need help to migrate the massive amount of data, and especially to review and recategorize it 1154362882 M * derjohn Hollow, yes ... 1154362895 M * Hollow and we should also install redirects for old pages, so google still finds them 1154362922 M * doener that will be the real PITA I guess... 1154362941 M * Hollow yep.. does anyone know if there is a list of all wiki pages? 1154362949 M * Hollow or do we need to extract this from db? 1154362982 M * doener Bertl could probably create a dump of the database, I'd volunteer to gather the needed data from it then 1154363015 J * meandtheshell ~markus@85-124-37-189.dynamic.xdsl-line.inode.at 1154363016 M * Hollow that would be great 1154363063 M * Hollow this is also the time to move to git ;) 1154363095 M * Hollow and another idea that came to my mind: offering svn access to vserver related projects, like tuil-vserver, vserver-utils, the web control panel etc 1154363151 M * doener svnview-a-like? (or whatever that web-frontend shit^Wstuff was called) 1154363283 M * Hollow well, i prefer trac 1154363307 M * daniel_hozac trac does seem awfully nice. 1154363317 M * daniel_hozac i especially like the interdiff support. 1154363328 M * Hollow yeah, and you can even disable the whole wiki, ticket etc stuff, to only get the nice svn browser ;) 1154363329 A * daniel_hozac uses viewvc currently. 1154363352 M * daniel_hozac trac is svn only though, no? 1154363357 M * Hollow yep 1154363368 M * daniel_hozac so moving to git would be kind of silly :) 1154363378 M * Hollow well, webgit for git, trac for svn ;) 1154363388 M * doener well, moving to svn would be braindead ;) 1154363398 M * doener (for Linux-VServer itself) 1154363405 M * Hollow yeah, git for kernel 1154363412 M * daniel_hozac yeah, i agree. but then it'd be easier to just use git for everything. 1154363416 M * Hollow and svn for the rest.. ;) 1154363423 M * Hollow would it? 1154363433 M * Hollow i like svn :) 1154363444 M * Hollow but, tbh, i don't care really 1154363448 M * Hollow as long as we don't use cvs 1154363449 M * Hollow :P 1154363453 M * daniel_hozac lol. 1154363474 M * daniel_hozac i still use CVS for my RPM sources :) 1154363487 M * Hollow cvs is broken by design, imo 1154363490 M * daniel_hozac (svn for real sources) 1154363508 M * daniel_hozac i like the individual file revision numbers though. 1154363557 M * Hollow we also need better PR ;) 1154363664 Q * coocoon Quit: KVIrc 3.2.0 'Realia' 1154363823 J * coocoon ~coocoon@p54A05797.dip.t-dialin.net 1154364111 Q * coocoon 1154364163 J * coocoon ~coocoon@p54A05797.dip.t-dialin.net 1154364643 M * Bertl doener, Hollow: yes, the contents of the wiki db should be no problem, it's quite simple, so we can dump just the latest version, or one version for each revision, etc 1154364663 M * Bertl it's a simple, structured, mysql database 1154365083 P * meandtheshell 1154366614 M * Hollow Bertl: great, can you upload the sqldump somewhere? or per email if it is not too big 1154366637 M * Bertl yes, will doso shortly... 1154366643 M * Hollow thx 1154367583 Q * renihs Ping timeout: 480 seconds 1154368035 J * waldi ~waldi@bblank.thinkmo.de 1154368039 M * waldi re 1154368387 J * meandtheshell ~markus@85-124-175-45.dynamic.xdsl-line.inode.at 1154370607 M * Bertl wb waldi! 1154371272 A * waldi just konverts the productive server to vserver from a pseudo-container in form of grsec 1154371537 P * _Roey Leaving 1154371716 Q * h01ger Quit: h01ger 1154371759 J * h01ger ~holger@socket.layer-acht.org 1154372852 J * pisco ~pampel@p50878A74.dip0.t-ipconnect.de 1154373501 M * Bertl okay,off for now, back later ... 1154373515 N * Bertl Bertl_oO 1154373772 J * mef ~mef@targe.CS.Princeton.EDU 1154373947 M * mef Just noticed that the latest version of FC5 has rpms for util-vserver-* available. 1154373960 M * mef Does this mean that the FC5 kernels have vserver merged in? 1154373970 M * mef Will check this out right now by pulling in the src rpm. 1154373982 M * mef Just curious whether someone here knew the answer off hand. 1154374118 M * daniel_hozac no. 1154374144 M * daniel_hozac Enrico's util-vserver packages in FE were copied with the other packages from fedora.us, i think. 1154374161 M * daniel_hozac (where vserver was used for the build system) 1154374167 M * mef Hi Daniel... 1154374172 M * mef glad to meet you. 1154374187 M * mef Someone from China pointed out that you hacked on the IPv6 support for vserver. 1154374193 M * mef Thanks for adding that functionality!!! 1154374205 M * daniel_hozac well, bonbons is doing the kernel stuff. 1154374226 M * mef Am playing with it and hope to integrate it on our PlanetLab systems (www.planet-lab.org), of which there are 600+ in 30+ countries. 1154374232 M * mef Oh 1154374233 M * mef I see 1154374243 M * mef bonbons: thank you! 1154374246 M * daniel_hozac i just did the userspace hacks. 1154374254 M * mef chbind6 1154374259 M * daniel_hozac right. 1154374288 M * mef so how will all of this pan out when a machine gets a new router advertisement with a new prefix? 1154374342 M * daniel_hozac i really don't know anything about IPv6... 1154374344 M * mef Should the guest get a new address so that it will also have the prefix. 1154374352 M * bonbons mef: it will be no different then a vanilla machine. 1154374353 M * mef daniel_hozac: that's funny. 1154374387 M * mef bonbons: I have not used v6 on a vserver based system. 1154374395 M * bonbons if you configured the auto-generated address for a guest, that guest will start being able to use it without further action 1154374403 M * mef bonbons: aha. 1154374413 M * mef excellent. 1154374428 M * mef So how do you pick an auto-configurable address for a guest? 1154374496 M * mef That is, one that is unique? 1154374517 M * bonbons as the addresses are autoconfigured as a concatenation of prefix and mac-address (with insertion of two bytes in mac address) those are predictable 1154374518 M * mef Or do you rely on DaD (dup addr. detection). 1154374549 M * mef I thought the lower 64 bits were the mac address + FFFE. 1154374557 M * mef And those addresses were for the host. 1154374581 M * mef Those of course are predictable. 1154374648 M * bonbons in vserver, you just say guest N can use ip address X (or set of addresses) 1154374707 M * mef I'll poke deeper by trying it myself. 1154374730 M * mef However, if you don't mind me asking some further dumb questions, where do you pick the unique IPv6 address X? 1154374735 M * bonbons for the auto-generated address, fffe is inserted in middle of mac address 1154374744 M * mef right, fffe is in the middle. 1154374748 M * daniel_hozac couldn't you just assign addresses which are unlikely to be used for automatic addresses? 1154374777 M * daniel_hozac such as 00:00:00 for the first three bytes. 1154374778 M * mef Are you assuming that the person assigning IP address can assign, um what's it called in IPv6 lingo, a subnet? 1154374813 M * mef daniel_hozac: you still need to pick a unique address among different physical hosts that are on the same network. 1154374815 M * daniel_hozac surely you can assign addresses in the same range as that of the host? 1154374866 M * daniel_hozac well, true, but that's also true for IPv4. 1154374867 M * mef bonbons: my assumption is that the host is assign a one or more prefixes and then has to contend with picking a unique lower 64 bit value. 1154374872 M * daniel_hozac so it's not really a new problem. 1154374894 M * mef daniel_hozac: automatic address selection was supposed to eliminate the problem that existed in IPv4. 1154374902 M * mef daniel_hozac: no need for dhcp, etc. 1154374913 M * daniel_hozac as i said, i really don't know anything about IPv6. 1154374922 M * bonbons mef: that's for the auto-configuration at host level, but in server environments you most often disable auto-configuration 1154374934 M * mef daniel_hozac: you simply pick the "unique" ether mac address as part of your unique IPv6 address and go. 1154374948 M * bonbons and assign the interfaces pre-defined addresses (part of your prefix) 1154374951 M * daniel_hozac mef: but what if that collides? 1154374974 M * mef daniel_hozac: if you have duplicate ethernet macs, you've got a whole different problem at hand. 1154374984 M * daniel_hozac mef: true. 1154375015 M * mef daniel_hozac: the vendors are all given a unique 48 bit prefix space and are not supposed to stamp out ethernet devices with the same mac address. 1154375020 M * daniel_hozac mef: so you want to implement some sort of s/fffe// for guests? 1154375053 M * mef daniel_hozac: well, yes, something like s/fffe/ would be one idea. 1154375055 M * bonbons as for vserver you do not assign an interface to a guest, but just one or more IP addresses autoconfiguration makes no sense 1154375071 M * mef But I have to think about bonbons response for server environments. 1154375111 M * daniel_hozac yeah, i'd say static assignments are what you want for servers... changing NIC shouldn't require a DNS update. 1154375127 M * daniel_hozac (that's assuming people don't memorize IPv6 addresses :)) 1154375137 M * mef bonbons: assuming one has dynamic DNS working properly/securely, one technically should not need to rely on statically assigned/picked IPv6 addresses for servers. 1154375167 M * bonbons daniel_hozac: that assumption is really fair for longer ipv6 prefixes or auto-generated IPv6 addresses! 1154375189 M * mef ok 1154375190 M * bonbons mef: that's true 1154375242 M * mef Any way, we have 600+ machines. It would be cleaner/nicer if I could rely on autogeneration + dynamic DNS rather than statically assigning the appropriate IPv6 addresses. 1154375268 M * bonbons but you still want to decide when a vserver should have ipv6 or not, thus it's the user-space tool's job to setup address 1154375336 M * bonbons so when you configure the guest's you give them ip addresses as you do for IPv4, or you could extend the tools to generate address based on presence of a flag 1154375369 M * mef I like the latter. generate a unique address based on presence flag. 1154375442 M * bonbons you could have a file /etc/vservers/interfaces/2/{ipv6,dev} and the util-vserver script building IP address on `cat .../dev` 1154375453 M * mef Ideally, a user-level tool would be notified when a new prefix is made available to the host and it then configures the guests with the appropriate auto-generated address along the lines of s/fffe/. 1154375501 M * mef This user-level tool would then also fix up /etc/hosts appropriately, while also reaching out to one or more dyn-DNS servers. 1154375507 M * bonbons yes, same way as it happens on vanilla machines for the dynamic DNS update 1154375524 M * mef do vanilla machines support dyn-DNS updates for IPv6? 1154375535 M * mef if so, sweet. 1154375540 M * mef Haven't looked into this for a while. 1154375567 M * mef As mentioned, I much prefer if servers did not use static IP addresses, but instead relied a bit more on DNS. 1154375590 M * mef While I know that DNS can be unreliable, there are folks who have built some highly reliable DNS services on PlanetLab. 1154375599 M * mef Any way... this is pretty good stuff. 1154375602 M * bonbons you need user-space tool for that, notification should be possible via netlink or the like, but never checked that 1154375713 M * mef cool 1154377338 J * insomnia1 ~insomniac@slackware.it 1154377338 Q * insomniac Read error: Connection reset by peer 1154377542 J * tatiane ~tatiane@201009036013.user.veloxzone.com.br 1154377568 Q * tatiane 1154377588 Q * Viper0482 Remote host closed the connection 1154377783 Q * zkbrsnie 1154378190 M * daniel_hozac isn't there a hotplug event? netlink would require a daemon... 1154378364 M * bonbons daniel_hozac: maybe 1154379383 J * mire ~mire@75-167-222-85.COOL.ADSL.VLine.verat.net 1154379433 P * pisc1 1154379898 Q * dna_ Quit: Verlassend 1154380441 Q * mef Quit: Download Gaim: http://gaim.sourceforge.net/ 1154380734 J * dos000 ~dos000@wsp05974758wss.cr.net.cable.rogers.com 1154380738 M * dos000 howdy 1154380875 M * dos000 i (once) used vtop on the host and i could see the processes with the guest names ... was i dreaming i cant get it to do that again. man is of no help as well 1154381135 Q * pisco Ping timeout: 480 seconds 1154381394 P * stefani I'm Parting (the water) 1154381962 Q * mire Quit: Leaving 1154381990 M * daniel_hozac dos000: vtop has never done that, vps does though. 1154382014 M * dos000 ooh ! .. thanks 1154383128 J * Aiken ~james@tooax6-125.dialup.optusnet.com.au 1154383766 A * Belu is away (ill be back later...) 1154383766 N * Belu Belu_zZz 1154384446 Q * coocoon Quit: KVIrc 3.2.0 'Realia' 1154384800 J * kir_home tis-844822@213.152.157.70 1154385892 Q * kir_home Quit: Ухожу я от вас 1154386195 Q * s0undt3ch Ping timeout: 480 seconds 1154387130 Q * dos000 Quit: Leaving 1154390298 J * shedi ~siggi@inferno.lhi.is