1111622723 M * ciphernaut cheers 1111624596 Q * ndim Ping timeout: 480 seconds 1111624684 Q * yarihm Quit: Leaving 1111624809 J * ndim hun@helena.bawue.de 1111625159 M * Bertl evening ndim! 1111626579 Q * no_maam Read error: Connection reset by peer 1111626714 T * Bertl http://linux-vserver.org/ | latest stable 1.2.10, devel 1.9.5, ng9.4 -- He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the wiki, and we'll forget about the minute ;) 1111626864 J * no_maam ~erik@datenzone.de 1111626877 M * Bertl evening no_maam! 1111628006 M * alexx great :) 1111628010 M * alexx (Release) 1111628037 M * Bertl thanks! 1111628056 M * alexx i can to the upgrade two fingers in the noze ? or i must be warning about something ? 1111628074 M * alexx (if i crash the prod server, i go on the road for 250 Km :/) 1111628077 M * Bertl upgrade from 1.2.10? ;) 1111628082 M * alexx nop 1111628085 M * alexx from 1.9.4 1111628093 M * Bertl should be pretty smooth ... 1111628097 M * alexx oki 1111628108 M * alexx if it's crash, you go to paris with me ? ;) 1111628120 M * Bertl if you pay the ride, sure ;) 1111628150 M * alexx héhé ;) 1111628170 M * alexx (sorry for accents) 1111628182 M * Bertl np, they look funny here ;) 1111628184 M * Doener hey, another release :) 1111628199 M * Bertl yeah, basically missed it, right? ;) 1111628211 M * alexx dl of 2.6.11.5 in progress ... 1111628251 M * alexx Bertl, have you more information about a <> fundation for vserver ? 1111628272 M * Doener yeah, also noticed my new "neighbour" (on 13thfloor) just a few minutes ago 1111628351 M * Bertl alexx: no breaking news there .. but I'm still in the 'collecting ideas' phase ... 1111628365 M * Bertl Doener: yeah, like that new company? 1111628369 M * Doener sure 1111628534 M * Doener regarding the ensc-needs-internet issue, it's s shame, that eastern germany got equipped with a glass fibre phone system, but (AFAIK) no company does really use that infrastructure to provide fast and affordable internet access... instead they improve technology to support dsl on glass fibre... 1111628570 M * Doener dark side of the "tech-haven" 1111628616 M * Bertl yeah, indeed ... 1111628659 M * Doener yeah, we got a super modern phone system, but unfortunately we're pushing older technologies that you can't use with that system. but hey, we're working on providing theoretically outdated technology to you, isn't it great? ;) 1111628833 M * Doener anyway, what i actually wanted to ask: i didn't have any look at 2.6.12-rc1 yet, any significant changes or simply no time for a look at it on your side either? 1111628883 M * Bertl no time here either ... but we can have a look right now ;) 1111628905 M * Doener no, i'm actually going to bed rsn ;) 1111628928 M * Bertl good idea ... guess I'll call it a day too ... 1111628938 M * Doener ok, good night then, Bertl! 1111628956 M * Bertl night Doener! night John Doe! ;) 1111628998 A * Bertl .o( memories of the Waltons ;) 1111629014 N * Doener Doener_zZz 1111629014 M * Bertl have a good night everyone! cya tomorrow! 1111629022 N * Bertl Bertl_zZ 1111629171 M * alexx ++ 1111629468 M * ciphernaut has someone added a xid feild for display in vtop? 1111629890 M * alexx i dont think 1111630094 Q * SVSwift Read error: Connection reset by peer 1111634521 Q * douglas Quit: 1111641933 J * douglas ~douglas@douglas.user.oftc.net 1111641984 M * douglas hey, I'm running 1.9.5 and for some reason when I start my vservers, /proc isn't mounted, is there a program I dont have running or something I'm forgetting to do? I upgraded from the stable version. 1111642336 M * douglas I'm also getting alot of warnings: ANSI does not permit the keyword 'inline' 1111643069 Q * douglas Quit: 1111648365 J * erwan_ho ~erwan@lns-vlq-39f-81-56-133-136.adsl.proxad.net 1111649997 M * ciphernaut have a good weekend all 1111650425 Q * erwan_ho Remote host closed the connection 1111654919 Q * jsambrook Quit: Download Gaim: http://gaim.sourceforge.net/ 1111655084 J * jsambrook ~jsambrook@aelfric.plus.com 1111655102 Q * jsambrook Quit: 1111657068 Q * SiD3WiNDR Ping timeout: 480 seconds 1111657114 J * SiD3WiNDR luser@bastard-operator.from-hell.be 1111657989 M * rs hi there 1111658055 M * alexx hi rs 1111658146 M * alexx 2.6.11.5-vs1.9.5: tested and adopted since 3:58 AM CET ;) 1111658165 M * rs nice :) 1111659757 J * jsambrook ~jsambrook@host-62-69-64-93.bsve.net 1111659999 M * TheSeer 2.6.11.5 1111660000 M * TheSeer wtf..? 1111660084 M * TheSeer wow.. quite a few fixes in there.. 1111660139 M * TheSeer so that means i better upgrade hmmm? 1111660150 Q * erwan_taf Remote host closed the connection 1111660261 M * rs TheSeer: the "Fix crash while reading /proc/net/route" is a little bit frightening 1111660297 M * TheSeer yeah.. that's what i figured ;> 1111660315 M * TheSeer oh well.. 1111660326 M * rs but it shouldn't really hurt for linux vserver installation while you hide this proc entry for users 1111660333 M * TheSeer do we have a final 1.9.5 patch against 2.6.11.5? 1111660339 M * rs yeah 1111660356 M * TheSeer k.. guess i'll update the fc3 howto then too ;> 1111661067 J * erwan_taf ~erwan@81.80.43.68 1111661665 M * TheSeer uhm.. how changed the fc3 howto and added this line? 1111661668 M * TheSeer #::: For the 2.6.xx kernel, maybe others, SELinux should be disabled as it conflicts with vserver, and an initrd should be built using the mkinitrd script for udev. 1111663506 M * TheSeer oum.. 1111663509 M * TheSeer Enable Legacy Networking Kernel API 1111663512 M * TheSeer y|n? 1111665552 Q * eyck Ping timeout: 480 seconds 1111667321 Q * grecea Remote host closed the connection 1111667581 J * grecea ~grecea@h-195-22-237-74.mdl.net 1111671920 J * Marimba ~das.brot@pD953F1D3.dip.t-dialin.net 1111672003 M * Marimba hello, can anyone help me with the eth0:0 problem on vservers while installing Confixx ? 1111672016 N * Doener_zZz Doener 1111672018 M * Doener morning! 1111672030 M * Marimba i am runing the alpha utils and kernel 2.6.11.5 1111672043 M * Marimba good morning Doener 1111672059 M * Marimba here it's 2_44 pm ^^ 1111672067 M * Doener same here ;) 1111672077 M * Marimba which country?! 1111672104 M * Marimba and why its morning for you ;) 1111672218 M * Doener germany... and it's morning because i just got up, who cares about the clock? ;) 1111672230 M * Marimba na dann sag ich mal guten Morgen ^^ 1111672245 A * Marimba kommt aus Brandenburg (Land / Stadt) 1111672285 M * Marimba du hast nicht zufällig eine Idee wie ich den vServer einstellen muss das ich Confixx installieren kann?! 1111672296 M * Doener in english please.... 1111672319 M * Marimba what the fuck ;) my english is so bad.. 1111672332 M * Doener IIRC confixx just takes the output from ifconfig to do it's stuff... easiest way to take care of that problem would be to replace ifconfig so that you just get the 'correct' output for confixx 1111672354 M * Doener s/it's/its/ 1111672418 M * Marimba ok i will try this - my other problem is settig the open file size limit on the vserver correctly 1111672590 M * Marimba and how i fake the ifconfig output - is there a faq for it?! 1111672652 M * Marimba just a file that echo the correct output from the root ?! 1111672834 M * Doener yep, just make it output something that confixx thinks is useful... the host's ifconfig output is a good starting point, replace the ip address to match your confixx registration data 1111672902 M * Marimba you know the answer for my other problem while set the ulimits ?! 1111672930 M * Marimba i set the open filesize limit for the root to 100000 then i set the limit for the vserver to 10000 1111672948 M * Marimba but the vserver than has 100000 :( 1111672962 M * Doener where did you set that limit? 1111673180 M * Marimba î am dooing it with the help of this page: http://ensc.de/tuc/util-vserver/doc/conf/configuration.html 1111673210 M * Marimba ./etc/vservers/vserver-name/rlimits 1111673271 M * Doener and how did you check the limit? 1111673276 M * Marimba ulimit -a 1111673279 M * Marimba in the vserver 1111673331 M * Doener ulimit vs. rlimit ;) the rlimits are per vserver and can be checked in /proc/virtual//limit 1111673369 M * Marimba the rlimits i set also 1111673373 M * Doener but AFAIK fsize is not really limited 1111673386 M * Doener Marimba: yes, you set those, but you checked the ulimits 1111673388 M * Marimba but the change was not visible on the vserver 1111673429 M * Marimba i see the rlimit change just here /proc/virtual/123432/stats 1111673459 M * Marimba my problem is if i am runing a server and get over 1024 connections the server stops responding 1111673466 M * Marimba and the very big problem 1111673472 M * Marimba the root hangs also 1111673481 M * Doener hm, i got no stats file there... 1111673494 M * Marimba and there is no chance to connect to my root 1111673494 M * Doener and status doesn't show any limits 1111673591 M * Marimba uhhm... 1111673614 M * Doener which patch version? 1111673627 M * Marimba alpha utils and kernel 2.6.11.5 1111673637 M * Marimba running on debian sarge 1111673777 M * Doener but which patch version? 1.9.5? 1111673789 M * Marimba i forget to ask if its possible to run a firewal on a vServer - at the moment i get a output tha i have no access 1111673828 M * Doener no, not possible yet, has to be done on the host. will be possible with ngnet (next generation networking) 1111673854 M * Marimba yes 1.9.5 1111674053 M * Doener hm, what kind of server is it that causes you problems? 1111674112 M * Marimba it's a server for applejuice - running on java 1.5 1111674145 M * Marimba (it's a german filesharing project) 1111674196 M * Marimba the ajserver runs without problems on the root - but let the system hangs if it's runing in my vServer 1111674214 M * Loki|muh which system? the host system? 1111674222 M * Doener yep... 1111674262 M * Marimba the host system is debian sarge - running on a Dual Xeon 2,8Ghz@4 GIG Memory 1111674269 M * Doener Marimba: did you limit memory for the vserver? 1111674281 M * Marimba no limits for the memory 1111674287 M * Doener hm... 1111674312 M * Marimba but i limit the memory from java using java -xmx1024m 1111674327 M * Marimba so tha java just can use 1 gig of it 1111674343 M * Marimba but the server never reach this limit... 1111674405 M * Marimba ahh fuck i just see i'm runing not 1.9.5 - 1.9.4 is my correct patch version 1111674461 M * Doener hm, where did you get the patch against 2.6.11.5? 1111674571 M * Marimba i have patched with the old patch 1111674584 M * Doener and that applied fine? 1111674588 M * Marimba yes 1111674592 M * Marimba it works for me 1111674605 N * Bertl_zZ Bertl 1111674610 M * Doener hm... 1111674620 M * Bertl fascinating ;) 1111674620 M * Doener morning Bertl 1111674625 M * Bertl morning folks! 1111674632 M * Marimba but i will update it to 1.9.5 now ;) 1111674656 M * Doener Bertl: didn't 2.6.11 introduce some changes we (you) had to take care of? 1111674712 M * Bertl yeah, there was something ... 1111674743 M * Doener ok, then i doubt that the patch applied as it should ;) 1111674764 M * Bertl well, let's wait and see ... probably the issue goes away with 1.9.5 ;) 1111674790 M * Marimba i stop my vservers and update now ;9 1111674806 M * Marimba hope my problems are gone than... 1111674843 M * Bertl if not, we will make them go away ;) 1111675003 M * Marimba if there are testsystems for you 1111675012 M * Marimba just try this in a vserver 1111675018 M * Marimba rm -Rf / 1111675034 M * Marimba a user from me was dooing it and crashes the whole server 1111675062 M * Marimba he was typing: rm -Rf / home/user 1111675077 Q * alexx Ping timeout: 480 seconds 1111675124 M * Bertl Marimba: sure, sec ... 1111675131 M * Marimba (a mistake from hime with the space after the / ) 1111675137 M * Marimba him 1111675212 M * Bertl chroot or namebased vserver? 1111675273 M * Marimba the normal vserver :) good question if it's chroot or namebased 1111675291 M * Marimba the servers are in /vservers/servername/ 1111675292 M * Bertl 1.9.x and with legacy config (one file) or new style config? 1111675298 M * Marimba new style 1111675348 M * Bertl okay, then it should be namespace based 1111675582 M * Bertl so on a chroot based (the first one I started ;) 1111675585 M * Bertl it gives: 1111675596 M * Bertl [root@XXXX root]# rm -Rf / 1111675597 M * Bertl rm: cannot chdir from `//dev/pts' to `0': Not a directory 1111675597 M * Bertl rm: `//proc/ide/hdc' changed dev/ino: Operation not permitted 1111675623 M * Marimba yes some errors but my root was crashed after that 1111675624 M * Bertl after that the running shell remains, but no files are left 1111675639 M * Marimba just try to restart the hostserver ;) 1111675643 M * Bertl host is still working fine 1111675651 M * Marimba that was the death for my root 1111675673 M * Bertl now let's see with namespaces ... 1111676565 M * Bertl again it works perfectly ... 1111676579 M * Marimba hmm... 1111676590 M * Bertl are you sure that your friend didn't do the rm on the host? 1111676607 M * Marimba no he has only his root pass 1111676641 M * Bertl well, this was with 2.6.11.5-vs1.9.5 now .. do you know what version you used? 1111676654 M * Marimba now its 1.9.5 1111676668 M * Marimba but it was 2.6.11.5-vs1.9.4 1111676680 M * rs re 1111676690 M * rs "morning" bertl :) 1111676700 M * Marimba patched with a patch for an older kernel... 1111676723 M * Bertl morning rs! 1111676741 M * Marimba it can be that this was the problem! 1111676799 M * Marimba on the new version i have problems again to set the right rlimit 1111676833 M * Marimba set it to 10.000 and get 100.000 (thats the change i made on the host for the open filesize limit) 1111676861 M * Bertl where do you set it? 1111676887 M * Marimba rlimits/nofile auf 10000 1111676891 M * Doener Marimba: your changes to the rlimits will never show up with "ulimit -a"... 1111676936 M * Marimba cat /proc/virtual/49155/limit show me that he has changed it right 1111676960 M * Marimba ulimit -a shows me the changes i made to the host 1111676967 M * Doener Marimba: ulimit != rlimit 1111676977 M * Marimba than how i set the ulimit?! 1111677022 M * Bertl how are your services started? 1111677034 M * Marimba i was thinking that rlimit replace the ulimits in 2.6 1111677051 M * Bertl yes, and the ulimits become what they are usually for 1111677072 M * Bertl ulimits can be changed inside a vserver (for example by pam) 1111677087 M * Bertl they are 'user' limits not per vserver limits 1111677110 M * Bertl (early versions abused those limits to make per vserver limits) 1111677126 M * Marimba what is pam?! 1111677132 M * Marimba i don't know it 1111677136 M * Bertl plugable authentication modules 1111677170 M * Marimba so i just must set the ulimit for the user my ajserver should run?! 1111677171 M * Bertl (see man pam for details) 1111677194 M * Bertl depends on what you're up to, if you want to have 'per vserver' limits, use the rlimits 1111677195 M * Marimba and if so - how i must do it... 1111677208 M * Marimba i want to have a limit per server 1111677234 M * Marimba i want give the host 100.000 at open file size limit and the vserver 10.000 1111677247 M * Bertl open file size? 1111677255 M * Marimba ulimit -n 1111677256 M * Bertl either file size, or open files ;) 1111677263 M * Bertl okay so open files then ;) 1111677277 M * Bertl just set the nofile limit to 10.000 1111677294 M * Bertl it will not show up in ulimit, but it will be there 1111677295 M * Marimba that ulimit -a shows me 100.000 in the vserver 1111677329 M * Bertl you can check /proc/virtual/*/limit for the limit and current usage ... 1111677345 M * Marimba yes of course this shows me the right limits 1111677368 M * Bertl but I take it as a feature suggestion that some folks might want to see those limits reflected in ulimit and friends ... 1111677400 M * Bertl (others might not want that, because some apps might depend on setting higher limits than they actually can use) 1111677424 M * Marimba if i have set the ulimit for the vServer - must i change it for the user should have more than 1024 also?! 1111677480 M * Marimba cause i changed the limit but if i am running the server and have too many connects (arround 1000) the vServer and the host stops responding 1111677488 M * Marimba and i don't know why 1111677497 M * Marimba no connect to ssh is possible than 1111677505 M * Marimba i have to reboot the mashine 1111677533 M * Marimba this my limits: 1111677535 M * Marimba PROC: 35 37 -1 0 1111677535 M * Marimba VM: 80951 84215 -1 0 1111677535 M * Marimba VML: 0 0 -1 0 1111677535 M * Marimba RSS: 22793 24812 -1 0 1111677535 M * Marimba ANON: 10495 10495 -1 0 1111677536 M * Marimba FILES: 371 386 50000 0 1111677536 M * Marimba OFD: 224 224 -1 0 1111677538 M * Marimba LOCKS: 3 3 -1 0 1111677538 M * Marimba SOCK: 85 85 -1 0 1111677540 M * Marimba MSGQ: 0 0 -1 0 1111677540 M * Marimba SHM: 181 181 -1 0 1111677576 M * Marimba ( output from cat /proc/virtual/49155/limit ) 1111677626 M * Marimba i am just wondering cause the server don't hang if runing under a user on the host - only if i start the server inside my vServer 1111677658 M * Marimba can it be that there problems using java inside a vServer?! 1111677725 M * Bertl well, there are always issues with java on any platform ;) 1111677743 M * Bertl but I don't think that your setup limits the java vm in any way 1111677774 M * Bertl do you have a test setup where you can do those 1000+ connects? 1111677819 M * Marimba no - i just start the ajserver and have to wait some hours 1111677838 M * Marimba there is a webpanel i can see how many users and connections i have 1111677851 M * Marimba if arround 1000 reached the server stop workin 1111677865 M * Marimba the only thing that runs is the ajserver than 1111677889 M * Marimba but the webpanel stops working to and the server don't let new users connect there 1111677900 M * Marimba but the old users can use it 1111677919 M * Marimba (users that connect before the limit reached) 1111677954 M * Bertl hmm, and this doesn't happen on the host? 1111677999 M * Marimba no there it works fine 1111678031 M * Marimba i have 4 ip's and can start 4 servers and running perfectly with more than 1500 users on each 1111678058 M * Bertl could you reproduce this situation where it 'hangs' locally and attach an strace to the java thingy? 1111678089 M * Marimba no idea how do this.. 1111678168 J * alexx ~alexx@82.225.136.176 1111678178 M * Marimba i will try it again with the newest java version 1111678223 M * Marimba and asking the programmer of the server building a server that makes loopback connections to reach the limit for testing 1111678349 J * CosmicRay ~jgoerzen@2002:4463:7269:1:20e:a6ff:fe66:c5a3 1111678480 M * Bertl Marimba: that would be a good idea ... 1111678481 J * sebd ~sebd@lesdeveloppementsdurables.org 1111678491 J * metaperl ~metaperl@nat.la.valueclick.com 1111678507 M * Bertl evening CosmicRay! sebd! metaperl! 1111678517 M * metaperl hi Bertl 1111678524 M * CosmicRay hello Bertl 1111678529 M * sebd hi* 1111678536 M * CosmicRay metaperl: Bertl is one of the developers behind vserver 1111678543 M * metaperl we are thinking of starting a cooperative internet host using linux-vserver 1111678547 M * metaperl I see 1111678582 J * Jason357 ~m00@wv-rndhll-cmts1a-a-105.shphwv.adelphia.net 1111678585 M * Bertl ah, interesting ... 1111678591 M * Bertl evening Jason357! 1111678595 M * Jason357 hello 1111678613 M * CosmicRay metaperl: what linux distro do you have now? 1111678626 M * metaperl hcoop.net runs debian 1111678642 M * CosmicRay do you have a workstation with debian on it that you could test stuff with? 1111678646 A * Jason357 has a need for a RHEL/clone server 1111678662 M * Jason357 CentOS, white box, etc 1111678675 M * CosmicRay well one of the nice things about vserver is that you can have, say, a debian master, and then virtual servers running debian, redhat, whatever 1111678676 M * Bertl what's the problem there? 1111678689 M * metaperl RHEL means what? 1111678698 M * CosmicRay metaperl: Red Hat Enterprise Linux 1111678700 M * Jason357 Red Hat Enterprise Linux 1111678704 M * metaperl oh 1111678716 M * CosmicRay there is probably little point to running RHEL in a vserver, though; aren't most of the things that make it "enterprise" simply kernel hacks? 1111678756 M * Jason357 the difference is support lifespan 1111678770 M * CosmicRay ah. 1111678783 M * Jason357 Red Hat/Fedora have I think an ~ 12-18month lifespam 1111678796 M * CosmicRay still, they will probably not support rhel in any kind of virtual server environment 1111678803 M * Jason357 while RHEL will be supported for severl years 1111678809 M * CosmicRay hell, debian goes twice that long between *releases* :-) 1111678811 Q * Marimba Quit: 1111678840 M * CosmicRay I've found Debian to make a very good OS for use in the virtual machine, and an adequate host OS 1111678842 M * Jason357 I run a specialized server application only supported in Red Hat 1111678849 M * CosmicRay actually I think that the vserver tools finally no longer suck in debian 1111678854 M * CosmicRay ugh 1111678874 M * CosmicRay I know there are people that run RH in the vserver, maybe ask on the list 1111678878 M * CosmicRay I've only ever tried it with debian 1111678884 M * CosmicRay and SuSE, which sucked 1111678901 J * j357 ~m00@wv-rndhll-cmts1a-a-105.shphwv.adelphia.net 1111678921 M * j357 *&*&^^& wifi 1111678926 M * j357 last I saw/said was: 1111678927 M * j357 CosmicRay I've found Debian to make a very good OS for use in the virtual machine, and an adequate host OS 1111678927 M * j357 Jason357 I run a specialized server application only supported in Red Hat 1111678927 M * j357 Jason357 I can keep it seperate though 1111678927 M * j357 * Disconnected (Connection reset by peer). 1111678945 M * CosmicRay I know there are people that run RH in the vserver, maybe ask on the list 1111678945 M * CosmicRay I've only ever tried it with debian 1111678945 M * CosmicRay and SuSE, which sucked 1111678969 M * CosmicRay suse sucked mainly because its installer and Yast made assumptions about the hardware 1111678989 M * daniel_hozac all my vservers and hosts are Fedora :) 1111679009 M * j357 I am not at all familiar with vserver, do you install an OS, and each vserver on the machine uses the same OS? 1111679033 M * CosmicRay well, it is more accurate to say each vserver uses the same kernel 1111679040 M * CosmicRay the userland can be different for each vserver 1111679045 M * metaperl daniel_hozac, what exactly is Fedora? 1111679061 M * j357 Fedora=RH 1111679064 M * daniel_hozac metaperl: fedora.redhat.com 1111679066 M * j357 more or less 1111679068 M * CosmicRay so, I could install a Debian OS on the machine.... with, say, kernel 2.6.9. Then I can install redhat, debian, whatever vservers. 1111679078 M * CosmicRay they all will be running under kernel 2.6.9 1111679081 M * CosmicRay but with their own userland 1111679085 M * j357 that is interesting 1111679086 M * j357 cool 1111679092 M * CosmicRay so they get their own init system, startup scripts, whatever 1111679107 M * CosmicRay the tricky part is the initial installation, since obviously you aren't partitioning disks and stuff like usual 1111679119 M * CosmicRay debian has this neat program called debootstrap that makes it real easy to do 1111679134 M * CosmicRay I know people do it with other dists too, but you'll have to ask them how :-) 1111679134 M * j357 can this be added on top an existing OS install? 1111679139 M * CosmicRay yes 1111679140 M * Bertl well, util-vserver creates vservers from network repositiories 1111679155 M * CosmicRay all you need is a kernel patch and a small tool suite 1111679162 M * CosmicRay from network repositories... of RPMs? 1111679169 M * Bertl http://linux-vserver.org/alpha+util-vserver 1111679169 M * daniel_hozac or debs. 1111679183 M * Bertl have a look at the examples 1111679192 Q * Jason357 Read error: Operation timed out 1111679196 M * metaperl if we want rackspace and bandwidth from someone, can we find someone using vserver? 1111679196 M * CosmicRay I see. nice. 1111679200 M * Bertl vserver build -m apt-rpm * -- -d fc1 1111679208 M * Bertl makes you an arp rpm fc1 vserver ;) 1111679219 M * CosmicRay metaperl: well, if you are looking for rackspace, you are presumably providing your own machine, so you don't have to bother with what they do 1111679228 M * CosmicRay metaperl: it is only an issue if they are providing the machine and os 1111679236 M * metaperl oh, I was using terminology I knew nothing about 1111679240 M * CosmicRay heh 1111679249 M * CosmicRay ok, well there are several ways to get hosting 1111679258 M * CosmicRay you can bring your own box 1111679276 M * CosmicRay they give you ethernet and power ports, you plug it in, and that's about where their responsibility ends 1111679281 M * j357 I think leasing a server is the way to go, complete with hardware support, bandwidth, and loaded with some OS 1111679296 M * CosmicRay then there are "dedicated server" options, where they provide and support the hardware 1111679304 M * CosmicRay sometimes they also provide and support the OS software 1111679319 M * CosmicRay these have a lower up-front cost, but may be more expensive monthly 1111679332 M * CosmicRay and usualyl don't have teribly impressive hardware specs 1111679352 M * CosmicRay the advantage is that they tend to have spare parts for just about everything on site, so if you lose a drive or a fan or something, they can have it back up quickly 1111679372 M * CosmicRay as opposed to you having to find parts and take it there, or fedex to the hosting facility and pay them to fix it for you, which can be a hassle 1111679373 A * j357 has AMD Athlon XP 3000+ 1GB RAM 2x 80GB HD 2TB bandwidth 1111679384 M * j357 from serverbeach.com 1111679411 M * CosmicRay that looks like a good price 1111679416 M * j357 support isnt great, they dont really do anything except fix hardware problems 1111679422 M * j357 I got a deal on it too 1111679424 M * aba well, the question is always: What do you need? "Good price" depends on your needs. 1111679425 M * CosmicRay and that's fine for me 1111679428 M * CosmicRay yup 1111679448 M * CosmicRay I've been maintaining debian servers for years. I don't want to pay anybody to maintain my OS, because I'd rather do it myself anyway 1111679455 M * CosmicRay so I want hardware support only in cases like this 1111679459 M * aba well, the cheapest ones here are about 20 Euro / month. 1111679476 M * metaperl I'm a Perl and Haskell programmer. I just want to read email, usenet and perhaps run an svn or darcs software repository 1111679485 M * metaperl j357 wants to run a jabber client 1111679490 M * j357 server 1111679492 M * j357 :) 1111679493 M * metaperl that's pretty much it 1111679501 M * CosmicRay you need rhel to run a jabber server? 1111679515 M * j357 Id like it to 1111679530 M * j357 I run one on this other server I speak of 1111679533 M * CosmicRay if you don't plan to add more people, I'd say just get the chepest miniserver you can from memset.com and split the cost 1111679546 M * CosmicRay if it's jsut the two of you, vserver is prboably overkill 1111679607 M * aba (and, if you want to use the servers really, you need serial console) 1111679615 M * metaperl we need about 40 people I think 1111679624 M * metaperl 2 people is not maxing out disk space our bandwidth 1111679633 M * metaperl but j thinks about 10-15 1111679647 M * CosmicRay well it may be if you get a 2GB server from memset that runs $35 a month or less :-) 1111679649 M * metaperl I think 10 root users 30 non-root 1111679655 M * aba metaperl: if you take a cheap box for 20euro/months or so, you shouldn't care about maxing things out. 1111679669 M * metaperl aba, like from where? 1111679672 M * CosmicRay aba: where exactly do you get a dedicated hardware hosting for $20EUR/mo? 1111679685 M * j357 he is talking about virtual 1111679690 M * j357 i think 1111679690 M * CosmicRay oh 1111679693 N * j357 jason357 1111679694 M * aba alturo.de 1111679705 M * aba jason357: no, about a real box. 1111679710 M * jason357 ? 1111679737 M * aba http://order.alturo.de/xml/order/Webserver;jsessionid=DF07B0C7F71F9CFC4A9CAC0B66E51968.TC12a?__frame=_top&__lf=Static 1111679741 M * metaperl their website is in german 1111679753 M * jason357 that is a bargain 1111679785 M * aba metaperl: yes, it's a german provider, the boxes are located in Germany. 1111679795 M * metaperl what is their up time like? and do they respond quickly to problems? 1111679801 M * aba (and you can reboot via http w/o additional costs) 1111679830 M * aba metaperl: uptime - no idea, but should be ok. Quick: usually yes, but only within bussiness hours. 1111679844 M * aba and: no extras, means: telephone support is quite expansive. 1111679855 M * metaperl j357, hcoop.net has a lot of people with ultra-minimal requirements. like a cgi guestbook for example... we could put some people on like that who have no need for root and save a lot of money 1111679902 M * metaperl maybe we should get off the vserver channel 1111679908 M * jason357 ok 1111679920 M * jason357 cooph? 1111679926 M * metaperl yes #cooph 1111679957 M * metaperl thanks for the link aba. I dont think they speak enough english in the case of an emergency 1111679984 P * CosmicRay Client exiting 1111680010 M * Bertl metaperl: you might also want to check: http://linux-vserver.org/VServer+Hosting 1111680024 M * metaperl ok will do 1111680398 M * kevinp Bertl: Why are there two different downloads of the same thing? linux-vserver-1.9.5.tar & patch-2.6.11.5-vs1.9.5.diff? 1111680417 M * kevinp Or are they different? 1111680444 M * Bertl well, the downloads are different, the result if you patch them is the same 1111680468 M * kevinp The tarball just has the patch in it right? 1111680486 M * Bertl the .tar contains all .diffs (which is currently just one) 1111680501 M * kevinp oh, ok. that makes sense 1111680510 M * Bertl so yeah, probably better to _not_ display the tar when there is just one patch ;) 1111680526 M * Snow-Man 2.6.12 will be out soonish, no? 1111680537 M * kevinp I'm getting ready to build a new server. I'll let you know how it works. 1111680547 M * Snow-Man Trying to decide if I want to try and hold off for 2.6.12 or noy. 1111680549 M * Snow-Man err, not 1111680582 M * kevinp 2.4 GHz, 2 GB RAM, raid 5, and a 9 GB scsi drive on a separate controller for swap 1111680608 M * kevinp should be good, right? 1111680633 M * Bertl Snow-Man: 2.6.12 will probably be less stable than 2.6.11.5, 'ya know that? 1111680660 M * Bertl kevinp: the raid 5 software or hardware? and how many disks? 1111680661 M * Snow-Man Bertl: It may also be better in some ways.. 1111680674 M * Snow-Man Bertl: I'm hoping the patches in 2.6.11.5 would be incorporated or better fixed in 2.6.123. 1111680677 M * kevinp Hardware, 4 - 36 GB disks 1111680679 M * Snow-Man err, 2.6.12, even. 1111680722 M * Bertl sure all 2.6.11.5 patches/fixes are part of 2.6.12 1111680732 M * Bertl but there are also new experimental features in there ... 1111680753 M * Bertl they will be fixed up in 2.6.12.x ... 1111680765 M * kevinp Do I want util-vserver-0.30.204.tar? Is that the latest utils or is there something else you were talking about that I should use? 1111680790 M * Bertl there is a cvs version, but the 204 tools should be fine for now ... 1111680802 M * Bertl probably enrico releases a 205 soon ... 1111680813 M * Bertl (then it would be advisable to upgrade) 1111680830 M * Snow-Man Bertl: It was my impression that there wouldn't be a patch for .x patch for 2.6.12 until there's a problem found. 1111680848 M * kevinp I will eventually migrate over a bunch of vservers from a 1.2 implementation, is there anything I should consider to make that better? 1111680853 M * Snow-Man Bertl: It didn't seem to me like the seperate .x tree was going to be doing things beyond being a subset of what goes into the next release. 1111680887 M * Bertl that's correct, but the difference is this: 1111680904 M * Bertl let's say development adds 50% bug fixes, and 50% new features 1111680916 M * Snow-Man I havn't actually had problems with the regular releases of late anyway.. 1111680920 M * Bertl then the 'sucker' tree will get 100% bugfixes 1111680931 M * Snow-Man These systems aren't really all that odd generally. 1111680937 M * Bertl (which are those 50% of the development ;) 1111680956 M * Bertl I agree, no problem with using bleeding edge stuff here ... 1111681080 M * kevinp What is the best how-to for installing the 2.6 patches? 1111681130 M * kevinp and configuring the kernel, etc 1111681136 J * Vudumen vudumen@perverz.hu 1111681224 M * Bertl evening Vudumen! 1111681231 M * Vudumen hi Bertl :) 1111681244 M * Bertl @all: are there any folks with 'unusual' graphical web browsers around? 1111681268 M * Bertl (i.e. something older or maybe something brand new) 1111681270 M * Vudumen currently not :( sometimes i use w3m on fb console with w3m-image extension 1111681275 M * Vudumen only links/lynx now 1111681282 M * kevinp Under development tools, are the rpms just parts of the util tarball? 1111681286 M * Bertl well, yes, I have them too ;) 1111681305 M * Bertl kevinp: no, the tar is the 'original' from enrico 1111681312 M * kevinp just mozilla and firefox right now 1111681321 M * Vudumen internet explorer if you are interested :) 1111681327 M * Bertl the rpms are for mandrake 10.1 but should work on other systems too ... 1111681331 M * Vudumen nowadays it's unusual :P 1111681333 M * Bertl Vudumen: sure ... 1111681337 M * Bertl http://212.16.62.58/ 1111681348 M * Bertl kevinp: yeah IE would be fine ;) 1111681362 M * Bertl other way round ;) 1111681385 M * Vudumen well with ie i see colours and style with firefox i don't 1111681386 M * Bertl firefox would be fine too ;) 1111681391 M * kevinp looks fine in firefox 1111681408 M * Bertl now the big question how does it look like ;) 1111681428 M * Bertl to simplify that, I'll upload a snapshot how it should look like ;) 1111681433 M * kevinp Looks like the content is slightly misaligned with the box above it, maybe one pixel to the left? 1111681550 M * kevinp http://roundsphere.com/Screenshot.png 1111681630 M * Vudumen http://inferno.interware.hu/ie.png and http://inferno.interware.hu/ff.png 1111681684 M * kevinp mine is firefox on linux 1111681699 M * Vudumen it's xp 1111681730 M * Bertl Vudumen: with firefox, could you check if you can _change_ the stylesheet? 1111681781 M * Vudumen hmm 1111681784 M * Vudumen with firefox now it's okay 1111681824 M * Vudumen with "webdeveloper" extension i can change 1111681825 M * Vudumen :) 1111681826 M * Bertl http://212.16.62.58/shot-01.png <-- here is how it is supposed to look like (and thanks a lot) 1111681877 M * Bertl no, I'm a liar ... wrong snapshot ;) 1111681913 M * Vudumen Bertl: now it's okay both with ie and firefox 1111681923 M * Vudumen well i go to drink a coffe. brb 1111681959 M * Bertl k, cya 1111681966 M * Doener Bertl: hm, maybe the infamous text/plain issue with the style in firefox? 1111682068 M * TheSeer i'd rather blame a broken webserver that then firefox 1111682078 M * TheSeer sending a css as text/plain is just bullshit 1111682081 M * Bertl well, yes, might be ... but if so, I really don't know how to solve it ;) 1111682093 M * TheSeer fix the mime.types 1111682096 M * Bertl TheSeer: any indication that this happens? 1111682098 M * Doener TheSeer: i didn't intent to blame firefox ;) 1111682120 M * TheSeer Bertl: load the css directly, view page info, check the mime type 1111682136 M * Bertl ah, you are the one who reported the test line, right? 1111682141 M * Doener TheSeer: did issue (at least on 13thfloor.at) only shows up sometimes... 1111682155 M * Bertl please could you paste your simple cmd-line test again? 1111682172 M * Doener Vudumen: please check your JavaScript Console in Firefox, it should issue a message if the stylesheet had the wrong mime-type 1111682176 M * TheSeer me testline? 1111682179 M * TheSeer no ;> 1111682190 M * Bertl damn who was it then ;) 1111682193 M * Doener Bertl: wget the stylesheet a few times 1111682222 A * Doener got no better test yet 1111682248 M * Bertl yeah, well, this is now the third apache .. and all do that ... 1111682250 M * kevinp Vudumen left for coffee... 1111682252 M * Doener ah, yes.. it is the infamous text/plain... 1111682254 M * Doener Error: The stylesheet http://212.16.62.58/style.css was not loaded because its MIME type, "text/plain", is not "text/css". 1111682268 M * Bertl (mime types contains the info, as well as an explicit apache entry) 1111682270 M * Doener got that on the second load with firefox, first one suceeded 1111682278 M * Doener stupid apache ;) 1111682310 M * Bertl well, it's rather annoying, but it's not just an apache bug, because the html now specifies the stylesheet type too 1111682344 M * Doener anyway, i'm off now, got to prepare the bbq, later folks! 1111682347 N * Doener Doener|gone 1111682364 M * Bertl cya 1111682383 M * Bertl TheSeer: any idea how to force apache to send the 'right' header? 1111682393 M * TheSeer hmm.. 1111682394 M * TheSeer wait a sec 1111682434 M * Doener|gone hardcore solution: have the stylesheet parsed by php and send a header yourself ;) 1111682445 M * TheSeer that's what i was thinking too ;> 1111682454 M * Bertl yeah, I fear that's the way to go ... 1111682475 M * TheSeer 1111682476 M * TheSeer ;> 1111682478 M * Bertl but I refuse to believe that apache 2.x is broken since big N releases 1111682497 M * TheSeer i'll test against apache 1.3.x 1111682555 M * kevinp There should be a file on your server like /etc/mime.types that defines that 1111682583 M * TheSeer kevinp: the point is that apache is LOOSING that information 1111682590 M * kevinp oh 1111682600 M * Bertl well, it is there, and there is also a /etc/httpd/conf/apache-mime.types 1111682601 M * TheSeer Bertl: i cannot reproduce that with 1.3.29 1111682644 M * Bertl *sigh* 1111682659 M * TheSeer i'll try a different box with 2.0.49 1111682749 M * Vudumen Doener|gone: yes it complains about the MIME type is text/plain 1111682765 M * TheSeer ttp://ox.salesemotion.de/style.css 1111682769 M * TheSeer http://ox.salesemotion.de/style.css 1111682776 M * TheSeer seems to work here 1111682794 M * TheSeer 2.0.52 1111682808 M * Bertl yep, seems so *deep sigh* 1111682832 M * TheSeer might be a 2.0.53 regression? 1111682844 M * Bertl do you have mod_mime_magic (not mod mime) activated? 1111682847 M * TheSeer or what version do you have? 1111682864 M * Bertl currently it's 2.0.50 1111682891 M * TheSeer grep "magic" httpd.conf 1111682891 M * TheSeer LoadModule mime_magic_module modules/mod_mime_magic.so 1111682901 M * TheSeer MIMEMagicFile conf/magic 1111682904 M * Bertl hmm, okay, let's load that one too ... 1111682951 M * Bertl nope, doesn't help ... 1111682963 M * TheSeer try updateing apache? 1111682982 J * j357 ~m00@wv-rndhll-cmts1a-a-105.shphwv.adelphia.net 1111682991 M * TheSeer the apache on that box is a distro version btw.. ( fedora core 3) 1111683000 M * Bertl well, I up and downgraded apache several times ... it didn't change this behaviour 1111683027 M * Bertl and what drives me nuts is that it happens randomly ... 1111683276 M * Bertl TheSeer: example: 1111683294 M * Bertl I add: AddType text/otto .css to the .htaccess file 1111683335 M * Bertl now I get alternating text/otto and text/plain :/ 1111683337 Q * jason357 Ping timeout: 480 seconds 1111683446 J * yarihm ~yarihm@80-218-3-115.dclient.hispeed.ch 1111683469 M * Bertl evening yarihm! 1111683474 M * yarihm eh bertl 1111683561 M * albeiro cafe del mar is good music for developing, you hardly feel passing time... 1111683565 M * albeiro ;p 1111683694 M * yarihm albeiro: as if you would feel passing time while coding anyway ,) 1111683722 M * albeiro heh, you are right. i do not code without music anyway ;p 1111684329 N * j357 jason357 1111684703 M * TheSeer Bertl: talked to some apache dev about it? 1111684712 M * TheSeer Bertl: since i cannot see any reason for it ;/ 1111684785 M * Bertl no, not yet, but probably will do, now compiling 2.0.53 ... 1111687624 J * douglas ~douglas@douglas.user.oftc.net 1111687638 M * Bertl evening douglas! 1111687664 M * douglas is there a sample config of the new way to do alpha vserver? (devel version) 1111687681 M * douglas hey bertl 1111687773 M * Bertl you know the flower page? 1111687785 M * Bertl you also know the alpha util-vserver page? 1111687807 M * Bertl you can get a quick example config by using the skeleton build method 1111687848 M * douglas http://linux-vserver.org/alpha+util-vserver 1111687853 M * douglas I know about that page. 1111687872 M * yarihm what's an alpha vserver? 1111687883 M * Bertl that's a confusing name ;) 1111687890 M * yarihm (just a vserver on linux/alpha?) 1111687908 M * Bertl yeah, that's right, but the tools (util-vserver) were called 'alpha' by enrico 1111687920 M * Bertl alpha similar to 'beta' 1111687931 M * yarihm ah ... that alpha ... 1111687975 M * douglas well thats what the pages is titled. alpha util-vserver so I assume thats what the vservers were called too, since everything has changed from stable, shouldn't we have a name for the "new" vserver way? so when newbies like me ask for help it doesn't confuse everyone else? :) 1111688082 M * Bertl yes, in an ideal world we would have a descriptive name for all this ;) 1111688107 M * douglas of course 1111688145 M * Bertl so we have to live with what's there ... 1111688160 M * Bertl actually the vservers did not change in any specific way ... just the config did ... 1111688186 M * Bertl our naming is now 'legacy vserver' for the old style vservers (and legacy config of course) 1111688224 M * douglas yea but thats a big diff on a hosting stand point. Which somewhere should be mention in the development webpage saying that the odl configs won't work or something, maybe I just missed it 1111688231 M * Bertl and the 'new' (new since a year or so) config is just called config and vservers are called guests ;) 1111688258 M * Bertl old configs work quite well, that's called backwards compatibility ;) 1111688385 M * douglas understandable. 1111688386 M * douglas btw 1111688418 M * douglas did anyone notice that the lastest alpha tools, now uses /usr/local/etc for all the vserver stuff instead of /etc/vservers like the previous versions of utils used? 1111688478 M * Bertl nope, this is a compile time option 1111688505 M * Bertl if you compile the tools with no --prefix or such, they use the 'default' path (being /usr/local in most distros) 1111688545 M * douglas true, but I've never had to actually use that option, ./configure used /etc/vservers 1111688557 M * douglas oh well its not that big of a deal :) 1111688602 M * douglas btw, even the alpha-vserver page uses /etc/vservers 1111688605 M * douglas :) 1111688655 M * TheSeer .oO/ rpmbuilld ( 1111688658 M * TheSeer ;) 1111688682 M * douglas hey bertl, whats teh flower page? 1111688822 M * Bertl http://ensc.de/tuc/util-vserver/doc/conf/configuration.html 1111688849 M * Bertl it's the config documentation for the 'new style' config 1111688892 M * douglas btw, who it gods name picked the background for that? lol 1111688913 M * douglas its not so much that its flowers, its the colors on the webpage that give me a headache. 1111688916 M * Bertl which one? 1111688969 M * Bertl ensc|w: you around? 1111688971 M * douglas the background being green, then dark green for some, then tan, then light green, then red, and the flowers too. 1111688972 M * douglas lol 1111688987 M * Bertl change the stylesheet ;) 1111689003 M * ensc Bertl: yes, but I will go shopping in a few minutes 1111689009 M * douglas naaa, I'd rather be like michael moore. 1111689029 M * douglas just complain alot and offer no solutions to the problems :) 1111689040 M * Bertl ensc: your html is incorrect ... 1111689061 M * douglas sorry to offend any of you michael moore fans. 1111689065 M * Bertl you should have just one rel="stylesheet" 1111689075 M * Bertl the other should be rel=alternate stylesheet" 1111689108 M * Bertl ensc|w: btw, I verified vkill for initpid processes, works with the 'real' ip 1111689124 M * Bertl I also queued a patch which will work for '1' (next release) 1111689313 M * Bertl vkill to xid (no process) will work too ... 1111689394 M * ensc thx, html should be fixed soon 1111689400 M * ensc regarding vkill... 1111689402 M * ensc +++ /usr/sbin/vserver-info 32 INITPID 1111689402 M * ensc ++ initpid=9292 1111689402 M * ensc ++ /usr/sbin/vkill -s INT -- 9292 1111689402 M * ensc vkill: vc_ctx_kill(): No such process 1111689403 Q * jsambrook Remote host closed the connection 1111689418 M * ensc ctx patch: yes (version vs1.9.5-rc1) 1111689476 M * Bertl are you sure that the process exists? 1111689489 M * Bertl because it might have terminated already .... 1111689583 M * ensc # /usr/sbin/vserver-info 32 INITPID 1111689583 M * ensc 7216 1111689583 M * ensc # vps ax | grep 7216 1111689583 M * ensc 7216 32 ftp ? Ss 0:00 /sbin/minit 1111689583 M * ensc ++ /usr/sbin/vkill -s INT -- 7216 1111689585 M * ensc vkill: vc_ctx_kill(): No such process 1111689620 M * ensc ok, back in 10min... 1111689626 M * Bertl ok, cya! 1111690925 M * ensc ok... back 1111690943 M * Bertl wb ensc! 1111690948 M * Bertl could you try soemthing for me? 1111690959 M * ensc what? 1111690972 M * Bertl you have my vkill tools somewhere? 1111691027 M * Bertl chcontext --xid 100 sleep 100 & 1111691051 M * Bertl # vps | grep sleep 1111691052 M * Bertl 471 100 tts/0 00:00:00 sleep 1111691063 M * Bertl # vkill -k 15 -p 471 -x 100 1111691075 M * Bertl terminates the sleep quite fine ... 1111691108 M * Bertl ahem, with initpid of course 1111691166 M * Bertl okay, give me some time, looks like I'm confused today ... 1111691300 M * Bertl # chcontext --ctx 100 --flag fakeinit sleep 100 & 1111691300 M * Bertl # New security context is 100 1111691305 M * Bertl # cat /proc/virtual/100/info 1111691305 M * Bertl ID:100 1111691305 M * Bertl Info:81195000 1111691305 M * Bertl Init:27 1111691311 M * Bertl # vkill -k 15 -p 27 -x 100 1111691320 M * Bertl [1] + Done(143) chcontext --ctx 100 --flag fakeinit sleep 100 1111691361 M * Bertl # chcontext --ctx 100 --flag fakeinit sleep 100 & 1111691365 M * Bertl # chcontext --xid 100 ps auxwww 1111691365 M * Bertl USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND 1111691365 M * Bertl root 1 0.3 0.7 2828 436 tts/0 S 19:08 0:00 sleep 100 1111691365 M * Bertl root 39 0.0 1.0 2532 668 tts/0 R 19:09 0:00 ps auxwww 1111691426 M * ensc ok, when I specify '--xid' it works 1111691451 M * Bertl hmm, you called it without xid? 1111691454 M * ensc yes 1111691471 M * Bertl hmm hmm, what semantic should we add to that? 1111691486 M * ensc send a signal to the process with pid PID 1111691487 M * ensc ? 1111691488 M * Bertl handle it like xid=1? 1111691542 M * ensc else, when 'vkill --xid 1' works, this would be fine also 1111691557 M * Bertl probably not ... 1111691580 M * Bertl is it a problem to get the xid of the context you want to kill processes in? 1111691591 M * ensc no, I know the context 1111691599 M * ensc it is just .... a little bit redundant 1111691609 M * Bertl okay, then this is the better way, maybe we separate pids in the future 1111691619 M * ensc see above: 1111691619 M * Bertl then every context could have a process with pid=27 1111691620 M * ensc 19:36 < ensc> +++ /usr/sbin/vserver-info 32 INITPID 1111691620 M * ensc 19:36 < ensc> ++ initpid=9292 1111691620 M * ensc 19:36 < ensc> ++ /usr/sbin/vkill -s INT -- 9292 1111691761 M * Bertl okay, in the future 1.9.6+ you can use pid=1 for the init process (in vkill) 1111691837 M * ensc ok, will be in .206+ 1111691880 M * Bertl feature-request for the tools: allow to create a config w for an existing server dir 1111691912 M * Bertl another feature request: allow to create vservers on symlinks or mounted partitions! 1111691927 M * Bertl (maybe all that stuff is already done ;) 1111692217 M * ensc the config stuff needs a more general solution 1111692227 M * ensc what do you mean with the second point? 1111692267 M * Bertl I had a few reports that if you do: 1111692291 M * Bertl mount /dev/vserver /mnt/vservers/XYZ 1111692304 M * Bertl vserver XYZ build .... 1111692320 M * Bertl that this moves away the original mount, which is not as intended ;) 1111692350 M * Bertl or let me put it the other way around, what is the 'correct' way to install an lvm based vserver? 1111692410 M * ensc dunno ;) I will probably allow empty directories in future versions... 1111692635 M * Bertl okay, great, or maye a special option ... 1111692678 M * Bertl because the one lvm (or whatever) mount for each vserver is very common ... if folks want to use quota stuff, for example 1111692697 M * Bertl what about --bind mounting a directory to /vservers/XYZ 1111692704 M * Bertl that would make sense too, or? 1111692797 M * jason357 can you meter bandwidth per vserver? 1111693000 Q * lilo Read error: Operation timed out 1111693030 M * Bertl jason357: which bandwidth? 1111693040 M * jason357 all 1111693491 M * Bertl there is some accounting done in /proc/virtual//cacct 1111693512 M * Bertl and you can setup some iptables accounting if you want to account specific network throughput 1111693852 J * lilo ~lilo@lilo.usercloak.oftc.net 1111694024 M * jason357 I was thinking give each person a vserver, make it simple and say you have 2 people sharing a host whihc gets 10 GB free bandwidth per month, so you tell them they get 5GB each... if 11 GB is used total, how do I know who to charge for the extra gig? 1111694102 M * Bertl in this case either split it according to the /proc/virtual//cacct usage or setup a precise packet logging by ip on the host 1111694181 M * jason357 how percise is /proc/virtual//cacct usage? 1111694234 M * jason357 does every virtual host require its own IP? 1111694299 M * Bertl virtual host = something used in httpd, virtual server (guest) is the term for the thingies running on a vserver host ;) 1111694338 M * Bertl and to answer your question, no not every vserver guest has to have a separate ip, but it helps 1111694357 M * jason357 gotcha 1111694378 M * jason357 without it, if you telnet to an IP, which machine answers? 1111694390 M * jason357 if multiple guests are on an IP 1111694473 M * Bertl yes, you can get around that, if you want to use a single ip by a simple trick 1111694494 M * Bertl you assign private ips to each guest, and do some fancy port->ip:port mapping on the host 1111694531 M * Bertl (for example first server, ssh:10022, second ssh:10122 1111694886 M * Bertl TheSeer: still around? 1111694891 M * TheSeer yes 1111694897 M * Bertl could you try something for me? 1111694905 J * duckx ~Duck@dyn-83-157-164-46.ppp.tiscali.fr 1111694908 M * TheSeer sure 1111694958 M * Bertl if I use lynx -head -dump http://212.16.62.58/style.css 1111694971 M * Bertl I always get text/otto (which is fine) 1111694994 M * TheSeer okay.. me too 1111695023 M * Bertl folks at the #apache channel don't know anything about such issues 1111695032 M * Bertl we already disabled mod_rewrite completely ... 1111695051 M * Bertl I also noticed that if you use wget, the first two runs are fine 1111695068 M * Bertl after that, you get text/plain ... after a pause, everything is fine again 1111695090 M * TheSeer strange.. 1111695307 M * TheSeer hmm.. so what do you want me to try? 1111695409 M * TheSeer i can verify that the 2nd wget call gets text/plain 1111695555 M * TheSeer interesting 1111695559 M * TheSeer is that traceable? 1111695591 M * TheSeer e.g. run apache in foreground nonforking mode and see what it does? 1111695610 M * Bertl I found it! 1111695611 M * TheSeer the thing is i cannot reproduce that on any of my apache installations 1111695620 M * Bertl the caching modules are broken it seems 1111695842 M * Bertl disabling mod_disk_cache seems to solve the issue ... 1111696015 Q * Doener|gone Quit: changing servers 1111696095 M * Bertl TheSeer: can you verify that the issue is gone now please? 1111696221 J * mikegrb ~michael@mikegrb.netop.oftc.net 1111696248 M * Bertl welcome mikegrb! 1111696251 J * Doener|gone doener@193.24.208.125 1111696254 J * Beirdo ~gjhurlbu@beirdo.usercloak.oftc.net 1111696256 M * mikegrb Beirdo: thanks 1111696259 M * mikegrb er Bertl 1111696272 M * Beirdo hah. caught you 1111696274 M * Bertl welcome Beirdo! 1111696283 M * Beirdo Thank you :) 1111696300 M * mikegrb Beirdo: see, I told you there was a channel on this network! 1111696314 M * mikegrb Beirdo: all the good channels are on oftc ;) 1111696314 M * Bertl a lot of oftc folks here? 1111696316 A * mikegrb ducks 1111696327 M * Beirdo So I'm trying to setup vservers on an ubuntu hoary host. I'm just confused as to the status for 2.6 kernels 1111696335 M * Beirdo mikegrb: not all of them, silly :) 1111696354 M * Bertl well, the 2.6 kernel is somewhat unstable, but this is about linux-vserver, right? 1111696369 M * Beirdo yes 1111696390 M * Bertl and I guess you can consider linux-vserver patches for 2.6 at least as stable as the kernel 1111696460 M * Beirdo OK, well, I haven't had any general stability problems yet 1111696505 M * Bertl want to know details about the features or so? 1111696506 M * Beirdo so, where would one normally find the 2.6 patches? I was looking last night but couldn't find anything. But the scotch might have had something to do with that 1111696517 M * Bertl linux-vserver.org 1111696559 M * Beirdo OK, I got that far :) 1111696572 M * Beirdo so the patches for 2.6 would be in VServer 1.9.5? 1111696595 M * Beirdo aha 1111696597 M * Beirdo there it is 1111696599 M * Bertl yes, 1.2.x is stable 2.4, 1.3.x is devel 2.4 (discontinued), 1.9.x is devel 2.6 1111696611 M * Beirdo Yeah I really should look for patches when more sober :) 1111696621 M * Bertl helps sometimes ... 1111696653 M * Beirdo something tells me this will be a bit of a pain to apply in a ubuntu tree... 1111696670 M * Beirdo wonder if the fine debian folks have done it already... 1111696759 M * Bertl well, don't do that, get a vanilla kernel 1111696769 M * Bertl (will save you a lot of troubles) 1111696787 M * Bertl debian is slowly getting on the right track, but it still needs work ... 1111696833 M * Beirdo Hmm. 1111696870 M * Beirdo OK, well, I can foresee this weekend getting hectic :) 1111696909 M * Bertl yeah? 1111696919 M * Beirdo there are some ubuntu/debian patches that I'll likely need in there for the system to work right. like using cramfs for the ramdisk, etc 1111696960 M * Beirdo so I'd have to try to make a ubuntu kernel patch from the vservers patch 1111696975 M * Beirdo I'd imagine most of said patch is for contexts? 1111696979 M * Bertl well, let me know if you need anything ... 1111696993 M * Beirdo OK. This should be a fun weekend. 1111697009 M * Beirdo is it known to work "OK" on AMD64? 1111697016 M * Bertl unbuntu uses debian kernels? 1111697024 M * Bertl or are they just 'similar' 1111697026 M * Beirdo ubuntu is an offshoot of debian 1111697054 M * Beirdo so they may have some extra patches, but are based on debian's kernels, AFAIK 1111697064 M * Bertl which kernel? 1111697093 M * Beirdo currently I'm using 2.6.10-4-amd64-generic 1111697116 M * Bertl okay, so debian kernel is 2.6.8-15 or 16 IIRC 1111697138 M * Bertl guess your chances with that one are much better ... 1111697163 M * Bertl and regarding x86_64 no issues there, just tools have to be 64bit too (if the kernel is 64bit) 1111697178 M * Beirdo OK 1111697191 M * Beirdo fair enough, I'd be building them anyways 1111697249 M * Bertl the 32bit compatibility layer is not done yet, I have no machine to test it on ;) 1111697283 M * Beirdo the linux-vserver-1.9.5.tar.gz is actually linux-vserver-1.9.5.tar.gz.gz, BTW. 1111697320 M * Beirdo Ah, Technical University of Vienna? 1111697321 M * Beirdo nice 1111697372 M * Beirdo that's right beside Karlsplatz station, no? 1111697454 M * Bertl yes, how did you get that now? 1111697477 M * Beirdo the university? 1111697482 M * Bertl yep 1111697488 M * Beirdo whois :) 1111697502 M * Beirdo and I've been to Vienna, I love that city 1111697521 M * Bertl i.c. btw, which url for the tar.gz.gz? 1111697530 M * Beirdo I might even have some pictures of your university. 1111697532 M * Beirdo http://www.13thfloor.at/vserver/d_rel26/v1.9.5/linux-vserver-1.9.5.tar.gz 1111697547 M * Beirdo I had to gunzip before tar tzvf 1111697588 M * Bertl interesting ... works fine here ... just tried it 1111697590 M * Beirdo now that might just be my browser being stupid, but that's what I saw :) 1111697631 M * Beirdo you like Vienna? :) 1111697673 M * Bertl well, I'm not much _in_ vienna, but it can be a nice city ... 1111697714 M * Beirdo Ah. A company I used to work for did some work for a company there, so I spent about 5 weeks in 2000 in Vienna 1111697764 M * Beirdo anyways. I'll try my luck with the patch when I get home, I guess. 1111697773 M * Beirdo are you on the channel often? 1111697820 M * Bertl yeah, relatively often ... 1111697834 M * Beirdo excellent. I'll try not to be a bother :) 1111698050 M * Bertl well, you do not have to try very hard, because I'm not easily bothered ... 1111698188 M * Beirdo ah, good :) 1111698409 J * DukeGangsta|aw ~uefgangst@p54869493.dip.t-dialin.net 1111698605 M * Bertl evening DukeGangsta|aw! 1111698614 M * DukeGangsta|aw hi ! 1111698614 M * Bertl wb lilo! btw ;) 1111698686 M * Beirdo heh. lilo's here too? :) 1111698690 M * Beirdo heya, lilo 1111698712 M * Beirdo I swear, I just keep crossing paths with the same set of people :) 1111698931 P * metaperl 1111698932 M * Bertl well, the world is not that small nowadays ;) 1111698946 M * Bertl s/small/large/ 1111698960 M * Bertl or s/not that/really/ 1111698991 M * Beirdo true, especially in the geek world :) 1111698996 Q * DukeGangsta|aw Quit: ( www.nnscript.de :: NoNameScript 3.81 :: www.XLhost.de ) 1111699038 Q * SiD3WiNDR Read error: Operation timed out 1111699045 J * SiD3WiNDR luser@bastard-operator.from-hell.be 1111699173 A * Bertl is a little confused today ... 1111699839 J * erwan_ho ~erwan@lns-vlq-39f-81-56-133-136.adsl.proxad.net 1111699982 M * Bertl welcome erwan! 1111700115 M * erwan_ho hey Bertl 1111700128 M * erwan_ho I saw 1.9.5 has been released, congrat' ! 1111700133 M * Bertl thanks! 1111700461 M * kevinp no thank you! It was perfect timing for a new release! 1111700589 M * Bertl well, you're welcome then! ;) 1111703364 J * rs_ ~rs@imhotep.rhapsodyk.net 1111703375 M * Bertl hmm, wb rs_ 1111703378 M * rs_ re 1111703389 M * rs_ Bertl: got a crash with the 2.6.11.5-vs1.9.5 1111703399 M * Bertl oops trace? 1111703409 M * rs_ soon available 1111703504 M * rs_ http://rs.rhapsodyk.net/vserver/oops-2.6.11.5-vs1.9.5.txt 1111703600 M * Bertl tx 1111703656 M * Bertl rs_: could you change the kernel config so that the BUG line is correct ... 1111703671 M * Bertl kernel folks consider it funny to disable that nowadays ;) 1111703690 M * Bertl (not for this oops, but for the next one ;) 1111703722 M * rs_ yeah ok 1111703728 M * rs_ what should I enable ? 1111703758 M * Bertl CONFIG_DEBUG_BUGVERBOSE=y 1111703771 M * rs_ ok 1111703783 M * rs_ is that new? because I never got this issue 1111703829 M * rs_ feel free to ask for addr resolutions 1111703853 M * SiD3WiNDR woop 1111703857 M * SiD3WiNDR I have 2.6.11.5 with 1.9.5 :s 1111703863 M * SiD3WiNDR but it's stable so far ;) 1111703874 M * rs_ SiD3WiNDR: could be related to our architecture 1111703906 M * rs_ there is some nfs things in the trace path 1111703914 M * Bertl rs_: yes, that's new as I said the option suddenly appeared and is off by default 1111703915 M * SiD3WiNDR ah, I don't use that on that machine :) 1111703938 M * Bertl it doesn't look nfs related 1111703993 M * Bertl I would need the following addresses: 80112d53,8011318c,8012ef50,80114d21, and 80112a78 1111704110 M * rs_ kernel/sched.c:654 1111704110 M * rs_ kernel/sched.c:1135 1111704110 M * rs_ kernel/wait.c:124 1111704110 M * rs_ kernel/sched.c:3051 1111704110 M * rs_ kernel/sched.c:590 1111704126 M * rs_ SiD3WiNDR: oups, you can start to be scared 1111704128 M * rs_ :P 1111704131 M * SiD3WiNDR lol 1111704148 M * SiD3WiNDR it's within reach to be rebooted, so not that bad :) 1111704208 M * rs_ Bertl: worth to be mentionned that there were many lines talking about killing some process because of reached memory limit 1111704234 M * rs_ ^MVM: killing process httpd 1111704235 M * rs_ ^MVM: killing process httpd 1111704236 M * rs_ ^MVM: killing process mysqld 1111704237 M * rs_ ... 1111704246 M * Bertl yeah, looks like we are hitting some error path here ... 1111704247 M * rs_ should be some missconfigured vservers 1111704921 M * Bertl rs_: kernel/sched.c:590 == 80112a78 ? 1111705091 M * rs_ yes 1111705132 M * Bertl ah, okay, I see ... 1111705160 M * Bertl that is a really healpful oops actually ... 1111705165 M * Bertl *helpful 1111705174 M * rs_ =) 1111705174 Q * lilo Quit: brb 1111705191 M * Bertl I give up with spelling for today ;) 1111705221 M * rs_ I won't blam you :) 1111705242 M * Beirdo time for ein beir 1111705260 M * Beirdo going home via the pub 1111705262 M * Beirdo :) 1111705536 Q * erwan_ho Remote host closed the connection 1111705998 J * lilo ~lilo@lilo.usercloak.oftc.net 1111706241 J * MrX freebsd@219.95.234.36 1111706459 M * Bertl evening MrX! 1111706913 M * Bertl rs_: btw, what browser do you use/have at ahnd? 1111706929 M * rs_ browser ? 1111706933 M * rs_ web browser ? 1111706936 M * Bertl yep 1111706948 M * rs_ safari of firefox depending on where I am 1111706964 M * rs_ why ? 1111706970 M * kevinp Bertl: I can check it again... 1111706998 M * Bertl rs_: up-to-date safari or older one? 1111707011 M * Bertl and do you have your IE still installed on the apple? 1111707018 M * rs_ depends one what you call up-to-date, it's the last stable 1111707031 M * rs_ yes I have 1111707044 M * Bertl could you check a page with the evil browser for me? 1111707051 M * Bertl (and upload a screenshot or so?) 1111707054 M * rs_ yeah sure 1111707067 M * rs_ by evil you mean IE mac I assume? :) 1111707142 M * Bertl you assume correctly ... http://212.16.62.58/ 1111707794 M * Doener|gone http://issues.apache.org/bugzilla/show_bug.cgi?id=30278 1111707798 N * Doener|gone Doener 1111707819 M * Doener text/plain 1111707836 M * Doener hm, i hit enter way too often... ;) 1111707878 M * Bertl Doener: text/plain should be gone 1111707901 M * Bertl I found and fixed it ... hopefully once and for all 1111707949 M * Doener ah damn, too late again ;) 1111707959 M * Bertl Doener: we have an issue: http://rs.rhapsodyk.net/vserver/oops-2.6.11.5-vs1.9.5.txt 1111707998 M * Bertl Doener: but thanks for looking into it! 1111708003 M * Bertl I really appreciate it! 1111708074 M * Doener any ideas regarding the bug yet? 1111708476 M * Bertl yeah ... it's an excellent trace IMHO ... 1111708504 M * rs_ I know I know, he he he :-> 1111708513 M * Bertl it looks like a task/process is put on hold (sched_hard) then get's killed by oom killer, which tries to wake it up (the hard way) 1111708569 M * Bertl of course, the task is on hold, and my check there triggers ... 1111708673 M * Bertl this brings up the question how to handle 'spontaneous' activation for tasks on hold ... 1111708699 M * Bertl guess we have to add some checks there which either put the task back on the runqueue or block the attempt 1111708732 M * Bertl rs_: hard_sched was enabled there, right? 1111708757 M * rs_ yeah of course 1111708768 Q * lilo Quit: bbiab 1111708768 M * rs_ do you want the kernel config ? 1111708779 M * Bertl no thanks! 1111708791 M * rs_ ok