1124755920 J * SiD3WiNDR luser@bastard-operator.from-hell.be 1124756238 Q * sladen Ping timeout: 480 seconds 1124756518 J * sladen paul@starsky.19inch.net 1124756549 M * Aiken Berlt have to go out, I'll try iscsi in a couple of hours 1124756576 M * Aiken in the meantime I have a vserver sitting on aoe and running and COW seems to work fine 1124756683 Q * Loki|muh Ping timeout: 480 seconds 1124756753 M * Bertl Aiken: excellent, thanks! 1124756776 M * Bertl evening SiD3WiNDR! 1124756793 J * daniel_hozac ~daniel@c-6f1472d5.010-230-73746f22.cust.bredbandsbolaget.se 1124756817 M * Bertl wb daniel_hozac! CDRW insertion? 1124756824 M * daniel_hozac thanks. 1124756834 M * daniel_hozac i had to move the CDRW drive to this computer. 1124756845 M * Bertl ah, the drive, that explains it :) 1124756850 M * daniel_hozac ;) 1124756993 J * Loki|muh loki@satanix.de 1124757046 J * keyser_soze ~cimarron@host89.201-252-9.telecom.net.ar 1124758094 Q * keyser_soze Quit: Abandonando 1124758113 J * keyser_soze ~cimarron@host89.201-252-9.telecom.net.ar 1124759732 Q * keyser_soze Quit: Abandonando 1124763797 J * aznboifaka ~brianlam9@dialup-4.248.4.197.Dial1.Honolulu1.Level3.net 1124763821 M * aznboifaka hello 1124763846 M * Bertl welcome aznboifaka! 1124763915 M * aznboifaka is this a help server? 1124763932 M * Bertl hmm, depends ... 1124763955 M * aznboifaka wow your name is in the green text thingy 1124763971 M * Bertl green text thingy? 1124763999 M * aznboifaka *** Now talking in #vserver 1124763999 M * aznboifaka *** Topic is 'http://linux-vserver.org/ | latest stable 2.0, 1.2.10, devel 2.1.0-pre1 -- 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 ;) ' 1124763999 M * aznboifaka *** Set by Bertl!~herbert@212.16.62.52 on Wed Aug 10 12:48:19 1124764011 M * aznboifaka see sent by bert1 1124764045 M * Bertl yeah, looks like I did set the topic :) 1124764097 M * aznboifaka are you teh admin here 1124764099 M * aznboifaka ? 1124764106 M * aznboifaka is everyone else idle? 1124764130 M * Bertl not much to administrate here ... 1124764157 M * Bertl guess some are idle/sleeping, others just waiting for something to happen ... 1124764174 M * aznboifaka oh looks like everyone is dead they don't leave the channel? 1124764178 M * aznboifaka i mean server 1124764211 M * Bertl some do leave, others just idle around ... 1124764219 M * aznboifaka y do ppl do that? 1124764238 M * Bertl what? 1124764245 M * aznboifaka just idle aroun 1124764248 M * aznboifaka just idle around 1124764292 M * Bertl guess some are interested to join if an interesting discussion is going on ... others might just hang around until their name is mentioned ... 1124764313 M * Bertl still others might read up what was written when they come back ... 1124764348 M * aznboifaka oh i see 1124764362 M * aznboifaka are you a progammer? 1124764370 M * Bertl now what brings you here, if I may ask? 1124764370 M * aznboifaka cuz thi si s a linux server 1124764388 M * aznboifaka idk i was just exploring around server to server 1124764404 M * Bertl yes I'm a programmer ... 1124764408 M * aznboifaka then i say linux theni thought someone might help me or teach me 1124764443 M * aznboifaka what comp programing languages do you know? 1124764469 M * aznboifaka c? c+? jave? python? php? etc 1124764472 M * Bertl quite a lot I guess (and even more if you count assembler) 1124764490 M * aznboifaka oh wow 1124764493 M * aznboifaka how old are you? 1124764515 M * Bertl I'm 35 1124764542 M * aznboifaka oh no wonder thoguht you were like my friend 15 cuz he's only 15 and he knows the ones i named 1124764567 M * aznboifaka aw man i gotta go do hw k c u man i'll drop in another time to ask more questions 1124764571 Q * aznboifaka Quit: 1124764587 M * Bertl k, cya :) 1124764629 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.13-rc6-vs2.1.0-pre4.diff 1124764645 T * Bertl http://linux-vserver.org/ | latest stable 2.0, 1.2.10, devel 2.1.0-pre4 -- 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 ;) 1124764725 M * Bertl okay, folks, I'm off to bed now ... have a nice whatever, cya tomorrow! 1124764758 N * Bertl Bertl_zZ 1124769885 J * stefani ~stefani@c-24-19-46-211.hsd1.wa.comcast.net 1124770570 J * keyser_soze ~cimarron@host89.201-252-9.telecom.net.ar 1124771477 Q * keyser_soze Quit: Abandonando 1124772384 M * Nicoli hey can someone tell me why i can't creatre vservers ? when i do i get this: 1124772385 M * Nicoli [root@vds1 etc]# vserver min-centos4 build -m yum --hostname domain.com --interface domain=eth0:192.168.0.136/24 --initstyle sysv \ 1124772386 M * Nicoli > --context 500 --force -- -d centos4 1124772388 M * Nicoli Renamed '/usr/local/etc/vservers/.defaults/vdirbase/min-centos4' to '/usr/local/etc/vservers/.defaults/vdirbase/min-centos4.~1124772232~' 1124772389 M * Nicoli Renamed '/usr/local/etc/vservers/min-centos4' to '/usr/local/etc/vservers/min-centos4.~1124772232~' 1124772391 M * Nicoli Renamed '/usr/local/etc/vservers/.defaults/vdirbase/.pkg/min-centos4' to '/usr/local/etc/vservers/.defaults/vdirbase/.pkg/min-centos4.~1124772232~' 1124772392 M * Nicoli cp: cannot stat `/usr/local/etc/vservers/.defaults/vdirbase/.pkg/min-centos4/yum/etc/yum.conf': No such file or directory 1124772394 M * Nicoli rm -rf /usr/local/etc/vservers/.defaults/vdirbase/min-centos4 /usr/local/etc/vservers/min-centos4 /usr/local/etc/vservers/.defaults/vdirbase/.pkg/min-centos4 1124773288 Q * stefani Quit: going away now... 1124774520 M * lilo hi Bertl(_zZ) 1124782518 J * menomc ~amery@200.75.27.13 1124782625 Q * mnemoc Ping timeout: 480 seconds 1124782868 J * sukria ~sukria@www.sukria.net 1124783360 J * Pazzo ~Pazzo@host130-250.pool8172.interbusiness.it 1124783365 M * Pazzo moinmoin! 1124786437 Q * Aiken Remote host closed the connection 1124786456 J * Aiken ~james@tooax6-089.dialup.optusnet.com.au 1124787700 Q * obi Ping timeout: 480 seconds 1124787723 J * obi ~obi@asus.saftware.de 1124790394 Q * renihs Ping timeout: 480 seconds 1124791102 Q * virtuoso Ping timeout: 480 seconds 1124791402 J * virtuoso ~s0t0na@80.253.205.251 1124791890 M * Nicoli anyone ere? 1124792331 J * Tenchi-san ~ask@208.53.159.170 1124792619 Q * Nicoli Ping timeout: 480 seconds 1124792795 Q * sukria Quit: leaving 1124793100 Q * Aiken Ping timeout: 480 seconds 1124796762 J * renihs ~renihs___@193.170.52.70 1124798284 J * mountie ~mountie@CPEdeaddeaddead-CM000a739acaa4.cpe.net.cable.rogers.com 1124799442 J * erwan_taf ~erwan@81.80.43.77 1124799449 P * erwan_taf 1124800260 Q * Hollow Remote host closed the connection 1124801165 N * Bertl_zZ Bertl 1124801200 M * Bertl morning folks! 1124801409 M * Tenchi-san mornin 1124801412 M * Tenchi-san how goes it > 1124801709 M * Bertl fine, thanks, Tenchi-san! 1124801713 M * Bertl and for you? 1124801716 M * Tenchi-san ok 1124801722 M * Tenchi-san reinstalled my server with fedora 4 1124801728 M * Tenchi-san hopeing i can get vserver to work this time 1124801749 M * Bertl heh, instead of? (please refresh my memory :) 1124801758 M * Tenchi-san centos 4.1 1124801772 M * Tenchi-san everythign seemd to ahve installed correcly but it just wouldnt create the vds's 1124801815 M * Bertl and, how is your luck with FC4 so far? 1124801851 M * Tenchi-san i finished runnign yum update 1124801852 M * Tenchi-san heh 1124801860 M * Bertl (and did you file some bugreports?) 1124801897 M * Tenchi-san i think it should work the sam with my centos instructions tho 1124801899 M * Tenchi-san same 1124801905 M * Tenchi-san i didn look 1124801970 M * Bertl the problem with the guest creation is that every distro has a slightly different setup and a completely different packaging system 1124801995 M * Tenchi-san i was jsut fallowing the howto step by step but i dunno how old it was 1124802005 M * Tenchi-san and it wanted to start with a debian vserv inside cent 4 1124802007 M * Tenchi-san so i dunno 1124802010 M * Bertl for example, you won't find a 'working' apt-get (for rpm) on any debian distro ... 1124802014 M * Tenchi-san im gonna try the last part sdiff this time 1124802063 M * Bertl usually debian is a good idea (as guest) especially woody or sarge worked for me (on mandrake and debian systems) 1124802097 M * Bertl but you are really not limited to the util-vserver tools with the guest creation ... 1124802143 M * Bertl you can, for example, create the guest system by installing whatever distro you want to use there, into a separate disk partition, then, copying that over and cleaning it a little up ... 1124802157 M * Bertl (that's how we installed the very first guests anyway) 1124802182 M * Tenchi-san srry im not very linux savvy 1124802189 M * Tenchi-san i know just enough to b dangerous heh 1124802198 M * Bertl lol 1124802620 M * Tenchi-san gah 1124802625 M * Tenchi-san this i dind expect 1124802626 M * Tenchi-san make[2]: *** [scripts/lxdialog/checklist.o] Error 1 1124802626 M * Tenchi-san make[1]: *** [menuconfig] Error 2 1124802626 M * Tenchi-san make: *** [menuconfig] Error 2 1124802638 M * Tenchi-san any ideas ? 1124802645 M * Bertl you are configuring the kernel? 1124802657 M * Tenchi-san im trying to 1124802665 J * Hollow ~Hollow@home.xnull.de 1124802676 M * Bertl you are missing some packages ...especially the cdialog-dev package 1124802680 M * Hollow heya 1124802692 M * Bertl Tenchi-san: whatever it is called on FC4) 1124802696 M * Tenchi-san gah, fc4 is realy turning out worse then cent0s 4.1 >_< 1124802704 M * Bertl hey Hollow! 1124802712 M * Tenchi-san ithink i shoukld just reinstall cent and try again from there 1124802715 M * Tenchi-san what do u think ? 1124802759 M * Bertl it's up to you, but if a missing package is able to stop you, I would not expect to get it working anytime soon :) 1124802779 M * Tenchi-san not i can fix that but if its gonna give me small probs jst in the beginning 1124802787 M * Tenchi-san i would suspect it will give me much larger probs lateron 1124802802 M * Tenchi-san wheresas cenot installed right the 1st time i tried 1124802806 M * Tenchi-san cent 1124802928 M * Tenchi-san or u think it will go the other way ? 1124802984 M * Bertl I heave no idea ... IMHO all distros need some tweaking to get it working as you want ti ... 1124803989 M * Bertl hmm .. network folks are improving the primary/secondary issue ... 1124803991 M * Bertl - /* 1. Deleting primary ifaddr forces deletion all secondaries */ 1124803991 M * Bertl + /* 1. Deleting primary ifaddr forces deletion all secondaries 1124803991 M * Bertl + * unless alias promotion is set 1124803991 M * Bertl + **/ 1124804183 M * Hollow Bertl: so, pre4 contains qorking context quota? :) 1124804229 M * Bertl not yet, but the quota hash abstraction to get it some testing ... 1124804264 M * Bertl although I doubt there will be quota specific testing at all :) 1124806512 J * oliwel ~mail-at-o@host-62-245-151-178.customer.m-online.net 1124806531 A * oliwel waves a hello to the crowd 1124806582 M * oliwel Hi Bertl -are you out there - this is the man with the Scheduler problem 1124806596 M * Bertl hey oliwel! 1124806608 M * oliwel Hi Bertl :) 1124806615 M * oliwel great to see you 1124806627 M * Bertl the pleasure is mine ... 1124806646 M * oliwel nice.. 1124806651 M * Bertl okay, first, let's get the testme.sh running and upload the output somewhere (e.g. pastebin.com) 1124806653 M * oliwel So -back to my problem... 1124806657 M * oliwel ok 1124806664 M * oliwel where can I get it ? 1124806695 M * Bertl http://vserver.13thfloor.at/Stuff/SCRIPT/ 1124806750 M * oliwel everything is green 1124806828 M * Bertl fine, could you upload it somewhere please? 1124806832 M * oliwel ok 1124806960 M * oliwel http://212.18.24.221/vs/test.out 1124807020 M * Bertl excellent, tx 1124807046 M * oliwel Its execellent software 1124807064 M * Bertl is there any reason why you are using 1.9.3/0.30.196? 1124807073 M * oliwel Fear :) 1124807082 M * Bertl (or just because you didn't update :) 1124807101 M * oliwel Its a procution maschine and I am currently working on the backup system 1124807118 M * oliwel So I dont want to crahs 90% of my customers :) 1124807120 M * Bertl okay, so it's running fine and you do not want to change a running system ... 1124807129 M * oliwel Not at the moment 1124807140 M * Bertl a reasonable explanation to me ... 1124807159 M * Bertl sec, digging out 1.9.3 :) 1124807165 M * oliwel I will setup the backup next week with up to date and if this passes test I will upgrade :) 1124807171 M * oliwel ok 1124807293 M * Bertl http://linux-vserver.org/ChangeLog26 1124807298 M * Bertl vs1.9.4 [2005.02.12] 1124807303 M * Bertl # fixed SCHED_PRIO w/o SCHED_HARD 1124807315 M * oliwel ok ... 1124807323 M * oliwel so sched_hard shpuld work? 1124807337 M * Bertl so, your kernel requires that you set sched_hard and have the HARD CPU enabled to make it work 1124807366 M * oliwel ok - so this is a bug in 1.9.3 and I should upgarde... 1124807386 M * Bertl you can check all the 'fixes' in the changelog up to the current stable 2.0 release 1124807413 M * oliwel I think I will finish the new server with current tools the end o the week 1124807423 M * Bertl please make sure to also update the tools when you upgrade the kernel to vs2.0 1124807445 M * oliwel yes of course - its a setup from scratch, so I use the current Gentoo Packs 1124807487 M * Bertl for now the sched_hard should work reasonably well .. I guess 1124807505 M * oliwel I have another question on RLimits...I cant figure out what I actually have to type in the rlimits directory for memory limits 1124807527 M * Bertl you mean the unit or what? 1124807535 M * oliwel Unit and format of the file, yep 1124807553 M * oliwel so I have a dir "rlimits" inside my vserver config - what now ? 1124807564 M * oliwel echo "1.5G" > data.hard??? 1124807575 M * Bertl units are pages, and it's just a number in the corresponding file (rss or as/vm/vml) 1124807597 M * oliwel So I must first find out the page size.... 1124807613 M * oliwel Can you gimme a hint what rss/as/vm are? 1124807613 M * Bertl so, for example echo "1024" >.../rlimits/rss 1124807640 M * Bertl would set the limits to 1024 pages (on x86 each 4k) 1124807654 M * oliwel What is "rss" ? 1124807658 M * Bertl rss = resident set size (means the pages actually in memory) 1124807664 M * oliwel aaah ok 1124807670 M * Bertl vm/as = virtual address space 1124807682 M * Bertl (that's what java can not get enough of :) 1124807698 M * oliwel so rss is kind of the physikal mem while vm/as is with swap 1124807707 M * Bertl and the vml is locked memory 1124807724 M * oliwel I start to undestand... 1124807726 M * Bertl oliwel: yes, it's very similar, though not identical 1124807740 M * oliwel Now - what is the "data" attrribue then ? 1124807792 M * Bertl for the rlimit it is not really used 1124807797 M * oliwel ok 1124807806 M * Bertl for the ulimit it limits the maximum size of a process's data segment 1124807851 M * Bertl (you can get an overview of the ulimits with man bash or similar) 1124807861 M * oliwel So if I put 250.000 in rlimits/rss the vserver can use a maximum of 1G of my RAM and it is no difference if this is physikal or swap ? 1124807872 M * oliwel I dont want to use ulimits 1124807897 M * Bertl the rss only applies to 'physical' memory, so it will not affect the swap 1124807916 M * Bertl your guest will be able to get up to 1GB of your RAM 1124807928 M * oliwel Hmm, I thought that memory managemnt ist done by the root server- is this wring ? 1124807945 M * oliwel and it can use as much swap as it wants ? 1124807950 M * Bertl it's done by the kernel actually, for all guests 1124807960 M * Bertl and yes, it can use as much as it wants 1124807967 M * oliwel can I limit swaptoo ? 1124807970 M * oliwel swap too... 1124807981 M * oliwel my space bar is a liitle bit laggy... 1124808014 M * Bertl not directly ... but a guest will not be able to use more swap space than VM/AS and will typically not use more than VM/AS-RSS 1124808035 M * oliwel ok seems reasonable - I will try this - thanks a lot 1124808045 M * Bertl you're welcome! 1124808087 M * oliwel BTW: I setup a nice project with vServer and will present this on our "OpenCA" workshop - so you can expect some new faces on the list mid october :) 1124808130 M * Bertl ah, sounds good! 1124808163 J * keyser_soze ~cimarron@200-55-65-102.dsl.prima.net.ar 1124808166 M * oliwel So - many thanks for the help, 1124808171 M * oliwel Have a nice week ;) 1124808179 M * Bertl thanks! you too! 1124808186 M * Bertl welcome keyser_soze! 1124808368 M * keyser_soze hello bertl 1124808725 Q * oliwel Quit: Chatzilla 0.9.68.5 [SUSE 1.0.4-1.1/20050511] 1124809032 J * mef ~mef@targe.CS.Princeton.EDU 1124809039 M * Bertl welcome mef! 1124809041 M * mef hey 1124809057 M * mef We finished upgrading to 2.0 and to the latest util-vserver tools. 1124809061 M * mef phew... 1124809064 Q * SNy Ping timeout: 480 seconds 1124809081 M * Bertl mef: excellent! 1124809085 M * mef vlimit question 1124809097 M * mef vlimit only appears to work when there is an active context. 1124809113 M * mef Do you guys normally use fake_init in your vservers? 1124809119 M * mef we don't. 1124809131 M * mef so there are some vservers that wont have procs. 1124809131 M * Bertl well, no context, nothing to limit ... 1124809143 M * mef ok 1124809146 M * Bertl nevertheless you can set it in the config ... 1124809157 M * mef that's what I figured I had to do ... 1124809159 M * Bertl this way it will be applied when the guest is started 1124809167 M * mef we let our users log in with a special "vsh" shell. 1124809200 M * mef This way we can let them ssh in and then confine them into a vserver. 1124809222 M * mef I suppose we are a bunch different from "normal" vserver, since we share a single IP across all vservers. 1124809240 M * mef So maybe you run sshd and the like directly inside a vserver, right?! 1124809242 M * Bertl hmm ... well if you like ... :) 1124809255 M * mef inside "each" vserver that is, right?! 1124809265 M * Bertl usually there are _many_ solutions to a single problem 1124809268 M * mef of course. 1124809272 M * mef We have a different one. 1124809280 M * Bertl most folks probably have one sshd for each guest ... 1124809305 M * mef I need to change our software to set the limits each time they log in via vsh. 1124809319 M * mef hmph... 1124809320 M * mef ok... 1124809323 M * mef I know what to do. 1124809332 M * mef thanks for letting me echo this off of you. 1124809346 M * Bertl mef: yes, your vsh should do all the stuff the utils do on startup 1124809359 M * mef will do 1124809361 M * mef thanks 1124809369 M * Bertl np, yw! 1124810761 M * menomc moin! 1124810790 M * Bertl morning menomc! 1124810801 N * menomc mnemoc 1124810809 M * mnemoc morning Bertl 1124810877 M * mnemoc Bertl: are vproc and vdlimit still addons part of last util-vserver? 1124810884 M * mnemoc s/still / 1124810888 M * mnemoc s/still // 1124810904 M * Bertl define addon? 1124810935 M * mnemoc optional tools 1124810972 M * Bertl they are part of util-vserver, so I'm not sure where you're heading to ... 1124810997 M * mef dumb question: which tools currently set all of the limits (cpu, mem, task, openfd). 1124811008 M * mef i.e., what uses the vlimit et. al. tools. 1124811044 M * Bertl maybe I should go to bed again ... somehow I do not understand a single question today :) 1124811051 M * mef or more precisely, which ones read the vserver conf file. 1124811068 M * mef sorry 1124811070 M * mef my questions suck 1124811074 M * Bertl the vserver tool reads all the config options 1124811082 M * mef gurgh. 1124811095 M * mef the vservers script... a'ight. 1124811118 M * Bertl vserver --help <--- that one :) 1124811145 M * Bertl well, actually it's a bunch of scripts doing the setup 1124811190 M * mnemoc Bertl: and cq-tools and vr-tools for 1.2 quota stuff? obsolete? 1124811226 M * Bertl mnemoc: you need them with/for the 1.2 quota patch (q0.14) 1124811259 M * mnemoc Bertl: ok, thanks :) 1124811295 M * Bertl IIRC, vrsetup was included some time ago ... 1124811328 M * Bertl yep, so the vr-tools should be obsolete by now ... but please verify that 1124811370 M * mef bertl: thanks 1124811375 M * mef looking into it right now. 1124811385 M * Bertl np 1124811409 M * mef We use vserver so differently from other folks. We are working on eliminating our dependency on util-vserver altogether. 1124811416 J * SNy 503222d81b@bmx-chemnitz.de 1124811438 M * Bertl wb SNy! 1124811452 M * SNy hm 1124811453 M * Bertl mef: you might want to look into Hollow's lib-vserver 1124811461 M * SNy my vserver host went down 1124811468 M * SNy again 1124811516 M * Bertl any logs or kernel traces? 1124811708 M * SNy my ISP doesn't tell me about that, unfortunately 1124811743 M * SNy I just see me ssh session die 1124811768 M * SNy and then sometime later I can login to the vserver with uptime 0 1124811772 M * SNy :/ 1124811786 M * Bertl anything in the logs? 1124811862 M * SNy of the vserver? 1124811863 M * SNy no 1124811913 M * Bertl hmm .. sec, you have a guest (vserver) or host with guests at you ISPs? 1124811954 M * SNy I just have a vserver there 1124811971 M * Bertl ah, okay, so they probably shut it down for whatever reason ... 1124811987 M * mnemoc Bertl: the build log, on a 9k tbz2 or 100k plain text? 1124811989 M * SNy I would appreciate it if they told me the reason 1124811994 M * SNy but alas 1124812003 M * SNy I am going to get me an uml soon 1124812008 M * SNy (no offense) 1124812021 M * Bertl mnemoc: 9k bz2 please 1124812054 M * Bertl SNy: none taken, UML is a nice piece of software ... but I guess the switch is not related ... 1124812067 M * SNy no 1124812069 M * SNy it's not 1124812071 M * SNy :) 1124812110 M * Bertl btw, why UML not XEN? 1124812120 M * mnemoc Bertl: sent 1124812414 M * SNy I know the people that run the uml hosting service 1124812421 M * SNy not personally 1124812437 M * SNy but I have better contact to them than to my current hoster 1124812471 Q * Hollow Remote host closed the connection 1124812477 M * Bertl SNy: i.c. well, maybe they want to start with linux-vserver too :) 1124812521 M * SNy I wouldn't count on it 1124812538 M * SNy also, I find vservers networking concept a tad strange 1124812553 M * SNy remember my complaint about 127.0.0.1? 1124812556 M * Bertl SNy: anyway, I would inform your ISP about the things you didn't like there ... so they have a chance to improve on that ... 1124812587 M * Bertl SNy: well, things will improve/change in this regard in the future ... 1124812620 M * Bertl (after all the networking is almost unchanged since more than two years now) 1124812655 M * SNy well, it's basically not a problem to be concerned about, once you are aware of it 1124812663 M * SNy and there comes the ISP into play 1124812666 M * SNy (again) 1124812695 M * SNy as you may guess by now, I am not too fond of them 1124812729 M * Bertl yes, nevertheless, I'd give them some feedback when you leave (or better right now) 1124812735 M * SNy their information policy is... something than can be greatly improved... to put it nicely 1124812849 M * Bertl a lot of providers (in the linux-vserver business) are very concerned about their curstomers ... some of them even refuse to update unless it's absolutely essential, because they do not want to disturb whatever the customer is doing ... 1124812857 J * Neubix ~brian@p54B072FE.dip.t-dialin.net 1124812863 M * Neubix Hi all 1124812869 M * Bertl hey Neubix! 1124812875 M * Neubix ey Bertl 1124812914 M * Neubix Medivh: you have found the script ? 1124812922 M * SNy can the ellipses be filled with a word there? 1124812930 M * SNy ;p 1124812943 M * Bertl SNy: which one :) 1124812945 M * SNy hehe 1124812948 M * SNy take a guess 1124812984 M * Bertl SNy: I guess yes *G* 1124812999 M * SNy don't get me wrong, it's more the tiny pieces that make it an unpleasant experience 1124813009 M * SNy not any major failure or something 1124813114 M * Bertl I just hope that you a) do not confuse provider issues with linux-vserver issues b) report back whatever issue you encounter to the appropriate places, and c) get happy with the UML guest ... 1124813155 M * Neubix hmm Medivh not alive ..:) 1124813272 M * SNy I am positive about a), b) and c) 1124814776 M * Bertl okay, I'm off for now .. back later 1124814781 N * Bertl Bertl_oO 1124814786 M * Neubix by 1124815094 J * Hollow ~Hollow@home.xnull.de 1124815419 Q * nox Quit: = If the world is without walls and fences -- who needs Windows and Gates ? 1124815481 M * Tenchi-san hey, is this bad ? 1124815482 M * Tenchi-san [root@vdss2 util-vserver-0.30.208]# /etc/init.d/vprocunhide restart 1124815482 M * Tenchi-san -bash: /etc/init.d/vprocunhide: No such file or directory 1124815781 M * mnemoc *g* 1124815783 M * renihs :) 1124815808 M * Tenchi-san the real wuation is, how important is vprocunhide 1124815824 M * renihs you will need it 1124815839 M * renihs in my humble opinion :) 1124815840 M * Tenchi-san is it a part of the util package ? 1124815876 M * Tenchi-san cause i jsut installed the util package and it didn give any errors 1124815888 M * renihs sys-cluster/util-vserver-0.30.208-r1 (/etc/init.d/vprocunhide) 1124815895 M * renihs belongs to util-vserver 1124815983 J * nox ~nox@noxlux.de 1124816007 M * Tenchi-san ok, what make install distrobution setting would i need for cent OS 4.1 ? 1124816064 M * renihs do u have /usr/lib/util-vserver/util-vserver-vars ? 1124816102 M * Tenchi-san no >_< 1124816120 M * renihs or util-vserver-vars somewhere? 1124816124 M * renihs that file *is* important 1124816138 M * Tenchi-san i dont even ahve a /usr/lib/util-vserver/ dir 1124816147 M * renihs somewhere else maybe? find it 1124816158 M * renihs if u just installed util-vserver check where u installed to 1124816202 M * renihs the vprocunhide init script isnt important, thats just a few lines, but the util-vserver-vars and the binaries etc :) 1124816274 M * Tenchi-san oops 1124816278 M * Tenchi-san its /usr/local/lib/util-vserver 1124816300 M * Tenchi-san so tis /usr/local/lib/util-vserver/util-vserver-vars 1124816303 M * Tenchi-san does exsist 1124816332 M * daniel_hozac check /usr/local/etc/init.d for vprocunhide 1124816338 M * renihs yah 1124816349 M * Tenchi-san yea its there 1124816369 M * renihs :) 1124816382 M * Tenchi-san hrm how can i change it so it confirms with where its normall supposed to be ? aka not in the /user/local dir ? 1124816391 M * renihs using configure --prefix=/bla is usefull :) 1124816395 M * Tenchi-san er installs where its normall supposed to be 1124816397 M * Tenchi-san ah 1124816401 M * Tenchi-san thx 1124816410 M * Tenchi-san if i reconfig and re make it shouldnt break anythign yea ? 1124816419 M * renihs no 1124816436 M * renihs but that question is ambiguous 1124816485 M * Tenchi-san im nto sure i fallow 1124816487 M * Tenchi-san not 1124816568 J * mess-mate ~mess-mate@lns-vlq-7-lil-82-254-193-118.adsl.proxad.net 1124816962 J * sim0n ~simon@p5089A2F9.dip0.t-ipconnect.de 1124817029 M * sim0n is it possible to have a loopback network int. inside a vserver ? (lo, 127.0.0.1) 1124817203 Q * Neubix Quit: Verlassend 1124818231 Q * sim0n Quit: Leaving 1124818342 J * Rob_ ~Rob@opencourse.org 1124820260 Q * duckx Ping timeout: 480 seconds 1124820275 Q * _ag_ Ping timeout: 481 seconds 1124820762 J * duckx ~Duck@mna75-1-81-57-39-234.fbx.proxad.net 1124821137 J * _ag_ ag@caladan.roxor.cx 1124821434 M * mess-mate simon: yes it can 1124822509 P * mess-mate 1124822919 J * neofutur_ ~neofutur@neofutur.net 1124823023 Q * neofutur Ping timeout: 480 seconds 1124823063 Q * mef Ping timeout: 480 seconds 1124825239 J * mess-mate ~mess-mate@lns-vlq-7-lil-82-254-193-118.adsl.proxad.net 1124825268 Q * dlp Ping timeout: 480 seconds 1124825753 N * mess-mate mess_o0 1124825779 N * mess_o0 mess-mate 1124825800 N * mess-mate mess_o0 1124825888 N * mess_o0 mess-mate 1124826928 M * Greek0 hmm, has anyone ever written something like a web frontend for vserver-stat and vserver start/stop/restart? 1124827626 M * Tenchi-san google open vps 1124827704 M * Greek0 ah, very cool. thanks 1124828498 Q * Tenchi-san Quit: ( www.nnscript.de :: NoNameScript 3.81 :: www.XLhost.de ) 1124829080 J * dejaVU ~cimarron@200-55-65-102.dsl.prima.net.ar 1124831593 J * yarihm ~yarihm@80-218-5-17.dclient.hispeed.ch 1124832580 Q * dejaVU Quit: Abandonando 1124833083 Q * keyser_soze Quit: Abandonando 1124833138 N * Bertl_oO Bertl 1124833149 M * Bertl evening folks! 1124833164 M * fobi hi Bertl 1124833177 M * Bertl hey fobi! how are you? 1124833340 M * Greek0 Bertl: ok, I think I got a patch for mips with !ABIO32. Looking at the include/asm-mips/unistd.h file it looks like N32/64 just use t0,t1 to get 6 registers to pass args in. 1124833380 M * Greek0 which are incidently also the regs that I use to setup the stack in the ABIO32 case. so just #ifdefing away the stack setup should work. 1124833407 M * Greek0 I've also made another minor change, I can send you patches tomorrow 1124833419 M * mugwump http://www.utsl.gen.nz/talks/vserver/ # anyone want to review / nitpick? 1124833743 M * Bertl Greek0: sounds good! 1124833750 M * Greek0 hmm, some minor details, slide 4d 1124833778 M * mugwump resource limits? 1124833797 M * Greek0 IMHO per-context quotas are a bit .. well not experimental. they work. they just work for the whole disk, including the host. 1124833812 M * Greek0 so you'd better only let that do the guests if you have them on a seperate partition 1124833863 Q * monrad Quit: Leaving 1124833868 M * Bertl mugwump: -ctx, now called -vs is funny :) 1124833873 M * mugwump right... so the CTX/UID split doesn't apply to that 1124833893 M * Greek0 the second point is that I don't really see how immulink fits under "resource limits". it fits well under sharing perhaps. but because "resource limits" is bold, it somehow looks that everything in that section is about resource limits 1124833902 M * mugwump Bertl: yes, because this is intended at a debian audience and the patch is still called -ctx in debian 1124833918 M * mugwump Greek0: point. 1124833958 M * Greek0 mugwump: typo in 6a "togethe r" in the title 1124833971 M * Bertl >100 releases, not too shabby! 1124833983 M * mugwump that's a guestimate, really 1124834083 M * Greek0 7e, why debootstrap a compilation area? 1124834087 M * Bertl regarding resource limits, you forgot the rlimits? 1124834129 M * mugwump Greek0: so main server is minimal 1124834134 M * Bertl (4d), and I agree with Greek0, ILI and now COW is not a resource limit 1124834154 M * mugwump rlimits? COW? :) 1124834164 M * Greek0 mugwump: uhm... ok 1124834197 M * mess-mate mugwump: IMHO your vserver..build...debootstrap example will never work in debian. 1124834200 J * keyser_soze ~cimarron@200-55-65-102.dsl.prima.net.ar 1124834207 M * Bertl Greek0: ad per-context quota, no they _are_ experimental ... 1124834225 M * Bertl Greek0: what you mean is called Disk Limits :) 1124834229 M * Greek0 btw, one thing I really like about vserver is that it somehow obsoletes LVM. at least one aspect where I thought lvm could be useful (for chroots) 1124834238 M * Greek0 Bertl: no, I was talking about quota support 1124834267 M * Bertl well, 2.6/vs2.0 has no context quota support 1124834281 M * Bertl Greek0: 2.1.0-pre4 has the first step for context quota 1124834297 M * Bertl on 2.4 there are the quota addon patches 1124834298 M * Greek0 Bertl: but you had that strange quota thing with vroot and stuff in 2.0-pre 1124834302 M * mugwump mess-mate: what's wrong with my example of using debootstrap? I tested that... 1124834307 M * Greek0 that's what I was talking about 1124834324 M * mugwump ah, rlimits is RSS limits, isn't it? 1124834333 M * Bertl Greek0: yes, vroot is there for a long time now and works quite fine ... 1124834348 M * Greek0 mess-mate: that's _not_ a problem of vserver. It's debians debootstrap not working for sid and possibly etch at the current time 1124834359 M * mugwump or Ubuntu breezy for that matter 1124834404 A * mugwump has 11 different debian distributions in vservers on his workstation 1124834413 M * Greek0 mugwump: you could mention that unification in some way makes LVM unnecessary for this purpose. 1124834424 M * mess-mate mugwump: sure ? on a sarge here with the 2.6.12.4 and the last debootstrap from unstable, it didn't work without the --resolve-deps option 1124834448 M * Greek0 mess-mate: do you have me on ignore or something? debootstrap from sid is _broken_ right now. 1124834468 M * Greek0 and I'm not even sure sarge's debootstrap can actually setup a sid system ATM. 1124834499 M * mugwump oh, rlimit is resource limit 1124834505 M * mugwump so how does that differ from ulimit? 1124834516 M * Bertl rlimits are for the entire context 1124834523 M * mugwump oh, magic 1124834524 M * Bertl ulimits are for each process 1124834550 M * mess-mate mugwum: here version 0.3.1.4 1124834589 M * Bertl mugwump: COW is copy on write, we have that in the current development prerelease (in combination with ILI) 1124834639 M * mess-mate mugwum: sorry, is from testing; 0.1.3.5 is from unstable :( 1124834644 M * mugwump ok, I remember having lengthy speculative discussions about how that might happen ... which method ended up working? 1124834668 M * mugwump mess-mate: is that a debootstrap option? My ubuntu doesn't have it 1124834691 M * Bertl mugwump: a modified/cleaned up? version from Jörn Engels mad-cow patches .. tailored to the ILI flag/system 1124834716 M * mugwump so, the inode number changes after a COW? 1124834720 M * mugwump after a write 1124834843 M * Bertl yep, first write operation breaks the link, makes a copy and returns that 1124834852 M * Bertl (first open with write intent) 1124834860 M * mugwump what about a chmod u-s ? 1124834870 M * mugwump That's what breaks libc upgrades with dpkg 1124834879 M * Bertl attribute changes are not allowed atm ...(on cow files) 1124834900 M * Bertl but we might get there sooner or later ... 1124834908 M * mugwump awesome 1124834984 M * Bertl regarding vunify (we now have vhashify, you know?) 1124834994 M * mess-mate mugwum:vserver test1 build -m debootstrap --interface test1=eth0:192.168.1.10/24 --hostname=test1.mydomain.org --context 42 -- -d testing -m ftp://ftp.debian.org/debian/ -- --resolve-deps 1124835017 M * Greek0 pretty nice work mugwump 1124835094 M * Bertl mugwump: what does 'mu' mean? 1124835132 M * Bertl mugwump: #11 please, please, please if you must mention VMware, also list QEMU 1124835272 M * Bertl mugwump: #13 we have no disk I/O limiting, but what do you miss regarding VM limits? 1124835311 M * mugwump well, I missed rlimits altogether 1124835345 M * Bertl okay, that probably explains it :) 1124835349 Q * yarihm Quit: Leaving 1124835363 M * mugwump mu is the correct answer to the question, "Have you stopped beating your wife yet?" 1124835395 M * mugwump it is a Japanese word alleged to 1124835395 M * mugwump mean "Your question cannot be answered because it depends on 1124835396 M * mugwump incorrect assumptions". 1124835428 M * Bertl aha, well, thanks for the information ... 1124835503 M * Bertl mugwump: #17 http://linux-vserver.org should be sufficient 1124835530 M * mugwump mu is certainly not in common English use, but it's in the Jargon file, for instance... 1124835539 M * mugwump ok I'll change that link and add QEmu etc 1124835601 M * Bertl btw, Aiken reported linux-vserver working on aoe, and NFS is supported too (server/client side) 1124835629 M * Bertl mugwump: pretty good overview ... 1124835636 M * mugwump excellent, so that NFS chbind conflict didn't affect the aoe driver 1124835657 M * Bertl the NFS chbind issue was resolved around 1.9.x 1124835677 M * Bertl (Lycos is using NFS filers :) 1124835747 Q * mess-mate Quit: next folks 1124835881 M * Greek0 Bertl: how simmilar are quota and dlimits actually? do you plan to merge those two sometime in the future? do you think it is possbile and/or makes sense? 1124835912 M * Bertl they are similar in where they are done/accounted ... 1124835933 M * Bertl (actually you will almost always find DLIMIT_* and DQUOT_* macros side by side) 1124835958 M * Greek0 mm, and some common low-level thingie that unifies this layer? 1124835962 M * Bertl but they are very different in _what_ they do and _how_ they are implemented 1124835977 M * mugwump so, dlimit's a catch-all for total size used by a vserver 1124835982 M * Greek0 I actually came up with this question by noticing that they are (nearly) always side by side :) 1124835990 M * mugwump whereas quota is about accounting users within a UNIX context 1124835993 M * Bertl Greek0: and unfortunately they also have different return values in case of error :) 1124836008 M * Bertl mugwump: yep 1124836037 M * mugwump so, has the XID/UID split been done for quota yet? 1124836045 M * Bertl Greek0, mugwump: if you dig into the older quota documents I did, you'll find that I originally planned to implement the 'context' quota as third quota type ... 1124836050 J * Aiken ~james@tooax7-008.dialup.optusnet.com.au 1124836059 M * Bertl morning Aiken! 1124836063 M * Aiken hi 1124836087 M * Aiken the iscsi initiator I normally use won't compile against 2.6.13 :( 1124836098 M * Greek0 "third"? 1124836100 M * Aiken so can not try iscsi with that kernel 1124836113 M * Bertl mugwump: xid tagging is part of _all_ vserver releases now ... it works for most filesystem (basically all which support 32bit uid/gids) 1124836138 M * Bertl Greek0: first is user, second is group quota ... 1124836172 M * mugwump ok, so can unmodified quota utilities work inside a vserver? 1124836174 M * Greek0 ic 1124836189 M * mugwump Or does it still need a helper of some kind? 1124836191 M * Bertl mugwump: for quota, it requires low level support because of the ugly way how linux quotas are implemented 1124836212 M * Bertl mugwump: no, using the vroot device you can use the quota tools to do everything you like 1124836243 M * Bertl mugwump: doesn't mean that you get user/group quota on a shared partition, this requires the quota patches (currently only available for 2.4) 1124836246 M * Greek0 but you'll set the quota for the whole disk/partition, i.e. also for the host using vroot 1124836274 M * Bertl yep, the vroot is just a proxy to make it secure ... 1124836339 M * mugwump ok, but if the quota file contains the full XID+UID, then it will only quota the user in a particular context, no? 1124836391 M * mugwump I think I'll just label per-context quotas as experimental for now 1124836428 M * Bertl mugwump: that was one of my previous ideas, but that doesn't work ... 1124836447 M * Greek0 ok, I'm off to bed now, good night 1124836460 M * mugwump nn! 1124836464 M * Bertl the problem here is that the tools insist on modifying the quota files manually 1124836467 M * Bertl Greek0: night! 1124836532 M * Bertl which has two implications: a) the files have to reside inside the chroot/namespace and b) guest root can modify any entries there ... 1124836652 M * mugwump virtualising that would be a royal PITA 1124836850 M * Bertl yep, that's why I had to figure a 'better' solution ... 1124836867 M * Bertl (which is currently implemented in the q0.14 patches) 1124836940 J * mess-mate ~mess-mate@lns-vlq-7-lil-82-254-193-118.adsl.proxad.net 1124836949 M * Bertl wb mess-mate! 1124836979 M * mess-mate Bertl: hi 1124837115 M * mess-mate Bertl: have changed the 'newserver' from debian for more options like for example --context. Is it usefull to determine that context when setting-up a vserver ? 1124837155 M * mugwump sheesh, does RLIMIT_CPU kill off a vserver entirely once it's used up it's allowance of CPU time? :) 1124837188 M * Bertl mugwump: it would if it was implemented :) 1124837234 M * mugwump are rlimits in 2.0? 1124837238 M * Bertl mess-mate: I really appreciate that you try to improve the setup, but patching newserver is like beating a dead horse ... 1124837251 M * Bertl mugwump: yep, they are 1124837305 M * mugwump what's not implemented other than RLIMIT_CPU (a probably dubious feature anyway), eg does the FD limit work? 1124837333 M * Bertl let me get a list of supported limits 1124837441 M * Bertl RSS, NPROC, NOFILE, MEMLOCK, AS, NSOCK, OPENFD, ANON, SHMEM 1124837505 M * mess-mate Bertl: finded at now newvserver the best way to set-up a vserver for debian. 1124837626 M * mess-mate Bertl: at least for a newbie.. 1124837647 M * Bertl mess-mate: well, if you think that creating guest configs which are more than a year old, do not support recent features and require legacy support, is the _best_ way for debian ... don't let me stop you ... 1124837735 M * mugwump I still image, anyway 1124838143 M * mess-mate Bertl: is version 0.1.10 from 13/5/2005 ? (debian/testing) modified now to the newest commands to setup a vserver.And is working :) 1124838178 M * Bertl great, does it create a new-style config? 1124838189 M * mess-mate Bertl: yep.. 1124838212 M * Bertl okay, let me see ... 1124838222 M * mess-mate Berl: would you give it a try ? 1124838227 M * Bertl sure ... 1124838242 M * mess-mate Bertl: send you now 1124838474 M * Aiken Bertl looks like I have viable alternatives to nfs, aoe http://pastebin.com/344504 and iscsi http://pastebin.com/344502 1124838534 M * Bertl mess-mate: http://vserver.13thfloor.at/Stuff/newvserver (please verify that this is the code you sent me) 1124838590 M * Bertl Aiken: looks good, but I suspect that once you start using restricted chbind areas, you will encounter issues ... 1124838608 M * Bertl Aiken: a simple but probably effective test would be: 1124838624 M * Bertl chbind --ip 127.0.0.1 # 1124838626 M * Bertl testfs.sh-0.06 -F ext3 -M /mnt/disk -D /dev/sdb1 -x 1124838635 M * Bertl chbind --ip 127.0.0.1 testfs.sh-0.06 -F ext3 -M /mnt/disk -D /dev/sdb1 -x 1124838648 M * mess-mate Bertl: yes it is. 1124838704 M * mess-mate Bertl: bad issues with chbind ? 1124838797 M * Aiken with iscsi everything passes chbind --ip 127.0.0.1 testfs.sh-0.06 -F ext3 -M /mnt/disk -D /dev/sdb1 -x 1124838816 M * mess-mate Bertl: why a chbind to the host localhost ? 1124838873 M * Bertl mess-mate: http://pastebin.com/344513 1124838925 M * Aiken same with aoe 1124838950 M * Aiken 2.6.12.4-vs2.0 1124838998 M * Bertl Aiken: well, that's good news ... now could you try to limit the 'other' end (server) with iptables or such to the host ip and/or examine the packets for guest ips? 1124839192 M * Aiken only allow the host ip access to the network block devices? 1124839235 M * Bertl yeah, something like that ... 1124839273 M * mess-mate Bertl: sended a command correction for etch on pastebin. I'll try it with sarge. 1124839292 M * Bertl mess-mate: http://pastebin.com/344523 1124839440 M * Bertl mess-mate: http://pastebin.com/344526 1124839527 M * Bertl mess-mate: btw, what is the purpose of the --domain (which seems quite different from the 'original yp domain) 1124839573 M * Bertl it looks like if you actually want to specify --hostname test105.test.org (at least to util-vserver) 1124839711 M * mess-mate Bertl: i've tested it for the 'sarge' dist and there are no problems ... . Well, a hostname for the vserver, don't it ? 1124839776 M * Bertl yes, but util-vserver has --hostname and --domain, but you specify linux-vserver.org as --hostname linux-vserver.org, not as --hostname linux-vserver --domain org, yes? 1124839805 M * Bertl because --domain is the yellow pages domain you want to use ... 1124839880 M * Bertl and it simply doesn't work at all ... 1124839929 M * mess-mate Bertl: setted also the localhost to 127.1.0.1 for further use with an own localhost. Then a little /network/interfaces and we can ping to our own localhost, some applications need it. 1124839975 M * Bertl mess-mate: okay, do what you have to do ... just don't bother me ... 1124840110 M * mess-mate Bertl: well, i followed the normal steps for a linux setup. Isn't the right way ? 1124840190 M * Bertl whatever you followed: a) it doesn't work, b) it's wrong in so many aspects ... 1124840460 M * mess-mate Bertl: are you on a debian distrib ? Thre is an advantage, i learned a bit the linux-vserver. So what other way to create a vserver ? 1124840481 M * Bertl yes, I tested the newvserver on a debian system 1124840497 M * Bertl cat /etc/debian_version 1124840497 M * Bertl 3.1 1124840525 M * mnemoc Bertl: what gcc/binutils/glibc you have there? 1124840546 M * Bertl gcc --version 1124840546 M * Bertl gcc (GCC) 3.3.5 (Debian 1:3.3.5-13) 1124840573 M * Bertl ii libc6 2.3.5-3 GNU C Library: Shared libraries and Timezone 1124840573 M * Bertl ii libc6-dev 2.3.5-3 GNU C Library: Development Libraries and Hea 1124840592 M * Bertl ld -v 1124840592 M * Bertl GNU ld version 2.15 1124840619 M * Bertl mnemoc: regarding your REG issues, can I get some account on your system? 1124840653 M * mnemoc it's a sandbox, but yes... send me your pubkey 1124840832 M * mess-mate Bertl: same for me. 1124841029 M * mess-mate Bertl: i've not the same glibc6 : Installed: 2.3.2.ds1-22 1124841076 M * mess-mate Sorry libc6 1124841139 M * Aiken running a vserver on iscsi, the only iscsi traffic is from the host ip 1124841153 M * Bertl Aiken: excellent news! 1124841320 M * mess-mate night folks 1124841326 M * Bertl night! 1124841334 Q * mess-mate Quit: leaving 1124841423 Q * duckx Ping timeout: 480 seconds 1124841455 J * duckx ~Duck@mna75-1-81-57-39-234.fbx.proxad.net 1124841575 Q * Doener Ping timeout: 480 seconds