1107141494 Q * gaber Ping timeout: 480 seconds 1107141740 J * gaber gaber@linux.gentoo.pl 1107143191 J * micah ~micah@mdm184-148.arc158.smfrct1.dasdial.com 1107143747 M * micah hi, I am trying to build the latest util-vserver tools (30.201) 1107143750 M * micah and I am getting an error: 1107143751 M * micah checking for ip... no 1107143751 M * micah configure: error: Can not find the 'ip' tool within '/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/X11R6/bin:/sbin:/usr/sbin:/usr/local/sbin' 1107143758 M * micah what is this "ip" tool? 1107144501 M * micah found it in iproute 1107144561 M * micah but now... configure: error: Can not find the 'vconfig' tool 1107144841 M * micah was in the debian vlan package 1107146722 M * lilo wishes: let me see if I have my faq alias on this client 1107146723 M * lilo I'm named after a favorite science fiction character, in John Varley's THE OPHIUCHI HOTLINE. I am *not* named after the LInux LOader, and I didn't write it! :) -- lilo, winter 1993 || Nor am I named after the little girl in the movie, nor the air mattress, nor Last In Last Out. ;) -- lilo, summer 2003 1107146728 M * lilo there we go 8) 1107147157 Q * nox Ping timeout: 480 seconds 1107147168 J * nox ~nox@c192045.adsl.hansenet.de 1107148301 M * micah i just tried to remove a vserver i created... I first removed /etc/vserver/ 1107148317 M * micah and then I tried to remove /vserver/, but I can't because some of the mount directories are busy 1107148418 M * micah I get things like: 1107148418 M * micah rm -rf tmp 1107148419 M * micah rm: cannot remove directory `tmp': Device or resource busy 1107150275 Q * micah Ping timeout: 480 seconds 1107160210 J * Val ~val@gj403.loria.fr 1107160228 M * Val Hi :) 1107161070 N * PazZzzzoooo Pazzo 1107161191 J * prae_ ~prae@ezoffice.mandrakesoft.com 1107161200 Q * prae_ Quit: 1107161213 Q * prae Remote host closed the connection 1107161272 J * prae ~prae@ezoffice.mandrakesoft.com 1107161345 N * Bertl_oO Bertl 1107161356 M * Bertl morning folks! 1107161395 J * BWare ~bware@212.26.196.41 1107162480 J * rs rs@ice.aspic.com 1107162490 M * rs hey dudes 1107162845 M * Bertl hey rs! how was your vaccation? 1107162945 M * rs was nice! not enough snow, but was fun thx :) 1107162975 M * Bertl you should have come here, we had about 40cm snow in the last week ;) 1107162995 M * Bertl and it's snowing right now ... 1107163010 M * rs hehe, yeah snow was everywhere in france too, except in you ski station... grr :) 1107163024 M * rs you/our 1107163055 M * Bertl ah, i.c. a special arrangement then *G* 1107163199 M * Bertl but I guess with a bunch of good friends the snow isn't that important ... 1107163223 J * mhepp ~mhepp@r72s22p13.home.nbox.cz 1107163240 M * Doener morning! 1107163243 M * Bertl welcome mhepp! 1107163248 M * Bertl morning Doener! 1107163257 M * rs hey Doener! 1107163263 M * Doener hi rs, Bertl! 1107164035 M * mhepp Bertl: Hi 1107164075 M * Bertl everything fine? 1107165264 M * Bertl hmm, guess fine but busy ;) 1107165588 M * Doener hm, should not run my x experiments on my on primary x server... 1107165619 M * Bertl ;) 1107165722 M * Bertl Doener: interested in a little ngnet hacking? 1107165780 M * Doener hm, the last time i tried to, i had to recognize that i lack too much knowledge in that area... guess i'll stick to crashing my x server for now ;) 1107165803 M * Bertl k, is fine for me ;) 1107166850 J * jsambrook ~jsambrook@aelfric.plus.com 1107166855 M * Bertl welcome jsambrook! 1107171433 Q * mhepp Ping timeout: 480 seconds 1107172935 J * mhepp ~mhepp@r72s22p13.home.nbox.cz 1107173655 Q * Seraph Quit: Changing server 1107173661 J * Seraph kk@projects.verfaction.de 1107175458 Q * berni Ping timeout: 480 seconds 1107175846 Q * BWare Read error: Connection reset by peer 1107176581 J * berni ~berni@2001:1b18:202::2 1107176626 M * Bertl wb berni! 1107176765 J * BWare ~bware@212.26.196.41 1107176779 M * Bertl wb BWare! 1107176842 M * matti ;] 1107176853 M * Doener mugwump: are you around? 1107176964 J * mbucc ~konversat@host-216-153-147-194.spr.choiceone.net 1107176980 M * Bertl wb mbucc! 1107176996 M * mbucc Bertl: greetings! 1107176996 M * Bertl hey matti! :] 1107177356 J * wfamy bill@lns-vlq-48-mar-82-251-25-178.adsl.proxad.net 1107177362 M * wfamy hi 1107177402 M * Bertl welcome wfamy! 1107177425 M * wfamy is any one have any link to explain how stoping vserver where daemon tools runing inside 1107177453 M * Bertl hmm, 'vserver stop' doesn't work? 1107177482 M * wfamy when i stop my vserver with fake init it can't stop svscan (surely because of init respawn) 1107177522 M * Bertl which kernel version, patches and tools? 1107177620 M * wfamy linux-2.6.9, patch-2.6.9-vs1.9.3.diff, util-vserver 0.30-14 1107177649 M * wfamy i try with the last utils but i had the same thing 1107177665 M * Bertl util-vserver 0.30.196 ? 1107177680 M * wfamy util-vserver-0.30.196 send me error too 1107177706 M * wfamy something about context error 1107177714 M * Bertl okay, could you upload the output of one of your stop attempts somewhere (e.g. pastebin.com)? 1107177751 J * micah ~micah@pool-64-222-118-112.burl.east.verizon.net 1107177803 M * Bertl welcome micah! 1107177850 M * wfamy done , http://pastebin.com/235699 1107177858 M * Bertl tx 1107177948 Q * mbucc Remote host closed the connection 1107178009 M * Bertl wfamy: could you redo that with export LC_LANG=C and export LC_ALL=C 1107178020 M * Bertl please? 1107178228 M * wfamy done http://pastebin.com/235701 1107178237 M * wfamy i add the mail8.conf 1107178240 M * Bertl tx 1107178278 M * Bertl JFMC: is that debian on the host? 1107178326 M * wfamy yes it is a sarge debian 1107178348 M * Bertl hmm, unrelated, but you should change the config to: 1107178358 M * wfamy i do not use debian patch for the kernel 1107178363 M * Bertl S_FLAGS="nproc fakeinit" 1107178372 M * Bertl ULIMIT="-HS -n 1024" 1107178380 M * Bertl S_CAPS="" 1107178393 M * Bertl and add an S_CONTEXT= 1107178444 M * Bertl anyway the fakeinit issue could be a kernel bug .. would you like to investigate it? 1107178619 M * wfamy try to stop with the new mail8.conf but same result. 1107178648 M * wfamy i am not able to investigate any kernel bug. 1107178650 M * Bertl yes, as I said, it's unrelated ... just that the config _should_ look like that ... 1107178664 M * Bertl wfamy: you are able, trust me ;) 1107178721 M * Bertl the question is, do you want to help identifying the issue (and probably fix it)? 1107178734 M * wfamy yes 1107178781 M * Bertl okay, first, do you know how the new config scheme works (you said you already tried the alpha tools, but with the old config, I guess?) 1107178937 M * wfamy i use a config i copy from old config (since almost 8 month) 1107178980 M * Bertl okay, and start/stop with the daemon tools worked before? 1107179008 M * wfamy i will have to go to sell piece of cake to get money to help tsunamy i be back in half an hour. 1107179038 M * Bertl okay, cya! 1107179273 M * micah goodmorning! 1107179306 M * micah I tried to remove a vserver, I wasn't sure how to do it properly so I removed /etc/vserver/ and then tried to remove /vservers/ 1107179310 M * micah but... 1107179322 M * micah rm: cannot remove directory `tmp': Device or resource busy 1107179322 M * micah etc. 1107179338 M * micah i am not sure how I can unmount those things, because they do not appear on the host system 1107179349 M * Bertl did you stop it before? 1107179408 M * micah yes, I believe I stopped it 1107179425 M * micah but... this was last night, and I am not positive this morning 1107179452 M * Bertl heh, okay, check with vserver-stat first 1107179470 M * micah everyone was asleep here last night :) 1107179529 M * Bertl unusual ;) 1107179545 M * micah its good to do once and a while :) 1107179553 M * micah ok, I do vserver-stat, and I get this error: 1107179556 M * micah WARNING: can not access /proc/uptime. Usually, this is caused by 1107179557 M * micah WARNING: can not access /proc/uptime. Usually, this is caused by 1107179563 M * micah (sorry for duplicate) 1107179573 M * micah and then there is a FAQ entry I should read 1107179588 M * Bertl hmm, and did you read it? 1107179598 M * micah I am reading now 1107179602 J * mbucc ~konversat@host-216-153-147-194.spr.choiceone.net 1107179639 M * micah Bertl: this is odd, because I am getting this in the host system, I am not inside the vserver itself 1107179652 M * micah and as root, I can read /proc/uptime with cat 1107179668 M * Bertl yes, but vserver-stat and vps run in xid=1 1107179685 M * micah ah ok, so I should run vprocunhide 1107179691 M * Bertl a badly configured proc security will cause that ... 1107179700 M * micah I have not configured any of proc security yet 1107179725 M * Bertl yeah, running vprocunhide (which is usually done on system startup) will do it for you ... 1107179736 M * micah ah ok, I have just installed things 1107179745 M * micah ok, vserver-stat now tells me: 1107179753 M * micah 0 88 286.2M 18.5K 2m23s29 1m27s78 20h21m03 root server 1107179754 M * micah 49152 5 10.6M 938 7h07m58 1h43m27 8h51m23 1107179770 M * Bertl hmm, yes that is probably the context you 'removed' ;) 1107179775 M * micah heheh 1107179797 M * Bertl so, best way is to do something like this: 1107179846 M * Bertl vnamespace -e 49152 vcontext --xid 49152 /bin/bash 1107179878 M * Bertl (after that you are almost inside the context) 1107179891 M * Bertl then check for processes with ps auxwww 1107179923 M * micah if I do that I get: 1107179924 M * micah Neither '--create' nor '--migrate specified; try '--help' for more information 1107179958 M * Bertl okay, use --migrate then 1107179976 M * Bertl vnamespace -e 49152 vcontext --xid 49152 --migrate /bin/bash 1107180033 M * micah ok, I've killed everything but my /bin/bash 1107180041 M * Bertl okay, don't leave it yet 1107180044 M * micah ok 1107180047 M * wfamy be back 1107180072 M * Bertl wfamy: wb! 1107180086 M * wfamy okay, and start/stop with the daemon tools worked before? yes it work for any vserver except those who use fakeini with daemontools 1107180137 M * Bertl micah: check for typical mount points like /proc or /tmp and /dev/pts and 'umount' them with the absolute and relative (to the vserver's root) 1107180173 M * Bertl wfamy: okay, so fakeinit + daemontools didn't work for you yet, right? 1107180187 M * wfamy yes 1107180207 M * Bertl okay, you have the 0.30.196 tools at hand? can you install them again? 1107180216 M * wfamy but never try without any procees follow by daemontools 1107180249 M * wfamy ok i remove debian utils and install 0.30.196 1107180306 M * micah Bertl: hmm, tried to do umount /tmp, and I get an error umount: /tmp: not mounted 1107180317 M * micah Bertl: but mount shows me all the filesystems that the host system has 1107180333 M * Bertl don't worry about that, just keep 'trying' to unmount all the possible places ... 1107180357 M * micah umount: devpts: not found 1107180357 M * micah umount: /dev/pts: must be superuser to umount 1107180383 M * Bertl /tmp /vserver//tmp ... 1107180412 M * micah so do: umount /tmp/vserver//tmp ? 1107180419 M * micah so do: umount /tmp /vserver//tmp ? 1107180423 M * Bertl no 1107180431 M * Bertl umount /tmp (you already did that) 1107180453 M * Bertl umount /vservers/karli/tmp (assumed your vserver was called karli) 1107180471 M * Bertl (or wherever your vservers are located ;) 1107180472 M * micah I do this inside the vserver still? 1107180477 M * Bertl yep 1107180490 M * micah I do that and it tells me I must be superuser to umount 1107180502 M * micah although I am (in the vserver) 1107180522 M * Bertl hmm, check with grep Cap /proc/self/status 1107180533 M * micah CapInh: 0000000000000000 1107180534 M * micah CapPrm: 00000000d44c04ff 1107180534 M * micah CapEff: 00000000d44c04ff 1107180560 J * _are_ ~are@212.185.72.226 1107180565 M * _are_ Hi 1107180574 M * Bertl okay, you have all necessary right ... just continue with the umounts 1107180577 M * micah hi _are_ 1107180581 M * _are_ did I mention nfs-user-server works within vserver? 1107180585 M * Bertl welcome _are_! 1107180617 M * _are_ if so, forget it, no locking avaliable. and no way for me to convince rpc.lockd to run so far. I guess I miss some CAP 1107180627 M * micah Bertl: hmm ok, so I seem not to be able to unmount anything 1107180648 M * Bertl okay, now try chroot /vservers/karli 1107180696 M * micah chroot: cannot run command `/bin/bash': No such file or directory 1107180724 M * Bertl hmm, yeah, you already removed it ;) 1107180728 M * micah heheh 1107180748 M * Bertl okay, I guess that is all you can do, now exit that bash and let's check if the dir still refuses to be removed 1107180767 M * micah that works! 1107180841 M * micah btw. I have a very small bug report, in vserver --help you see: 1107180841 M * micah running ... succeeds iff the vserver is running 1107180848 M * micah iff should only be "if" :) 1107180858 M * Bertl great, and next time, please shutdown your vserver before removal ;) 1107180873 M * micah Bertl: is there a better removal process that I should use? 1107180911 M * Bertl micah: bug-report on the ml and Enrico, and feel free to suggest a vserver destroy/delete/remove ... I'm missing something like that too ... 1107181098 J * berni_ ~berni@2001:1b18:202::2 1107181118 Q * berni Read error: Connection reset by peer 1107181143 M * Bertl wfamy: just let me know when you have the alpha tools running ... 1107181177 M * micah Bertl: done, thanks :) 1107181187 M * wfamy i have to change the vrootdir and i recompile 1107181193 M * Bertl micah: you're welcome, and thank you! 1107181237 M * micah I am using util-vserver-0.30.201 now, I feel very cutting edge :) 1107181280 M * Bertl hmm, enrico made a new release? 1107181327 M * micah Bertl: I think it was dated jan. 27th 1107181341 M * micah Bertl: I could not compile the other one because of the dietlibc version problem 1107181348 M * Bertl yep, seems I missed it completely ;) 1107181368 M * Bertl thanks for the info ... ;) 1107181379 M * micah no problem 1107181403 M * wfamy vserver-stat now give me ctx intead of name 1107181427 M * Bertl wfamy: that's fine ... 1107181447 M * Bertl we will now try to create a new-style config for your existing vserver ... 1107181453 M * micah Bertl: i noticed when I was compiling this release that I had some problems because I did not have "ip" or "vconfig", so I had to figure out what these were. Was there a README somewhere that I missed listing the dependencies needed? 1107181464 M * wfamy ok 1107181513 M * Bertl wfamy: your guest is debian too? 1107181570 M * wfamy all debian 1107181629 M * micah wfamy: are you using the debian packages? 1107181637 J * lars||| ~ledj@213-156-52-105.fastres.net 1107181664 M * Bertl welcome lars|||! 1107181683 M * lars||| now is that an automated message or not? 1107181683 M * lars||| :) 1107181693 M * micah I dont think it is :) 1107181698 M * lars||| ahh ok 1107181708 M * lars||| first time here, so I wouldn't know :) 1107181726 M * micah Bertl is just very welcoming to people :) 1107181740 M * wfamy micah: i use debian package with my home patch kernel without any problem since last week where i canot stop vserver with daemon tools inside 1107181816 M * lars||| I see 1107181838 M * lars||| well, reason I am here is that I ran out of luck reading the docs 1107181844 M * Bertl wfamy: try 'vserver mail8n build -m skeleton --hostname mail8 --interface eth1:192.168.0.108/24 --context 42 --flags fakeinit -- -d sarge' 1107181896 M * Bertl lars|||: hmm ... EOOLUCK? 1107181904 M * micah hehe 1107181931 M * wfamy nothing happend 1107181963 M * Bertl well, are you sure? I'd expect a new directory called mail8n and a config mail8n with an almost empty dir 1107181992 M * Bertl (wherever your vservers/configs are located) 1107182023 M * wfamy ok 1107182098 M * micah I installed the util-vserver alpha tools in /usr/local and I see that there are startup scripts put in /usr/local/etc/init.d -- how many of these do I want to make sure start up on system boot? I do not want to have the v_ scripts start because I want to bind processes to certain IPs in their config (like sshd_config). It seems I should make sure vprocunhide and vservers-default, but I do not know rebootmgr, vservers and vservers-le 1107182098 M * lars||| bbl' 1107182134 M * Bertl if so, I would ask you to copy over the contents of you mail8 server to the mail8n dir, except for the /dev entries ... 1107182141 A * lars||| is away: gone 1107182155 M * Bertl cya 1107182514 T * Bertl http://linux-vserver.org/ | latest stable 1.2.10, devel 1.9.{3,4-rc4}, ng8.13 1107182552 M * micah Bertl: should I have rebootmgr, vprocunhide, vservers, vservers-default and vservers-legacy all start at boot? 1107182572 M * Bertl rebootmgr should not be part of any tools anymore 1107182586 M * micah ok 1107182614 M * Bertl don't know about vservers, but vprocunhide, and vservers-default and/or vservers-legacy should be enabled if you want to autostart the vservers 1107182637 M * micah legacy is if I had old 1.2 vservers? 1107182651 M * Bertl yep, precisely (with the one file config) 1107182653 M * wfamy ok rsync finish everything exept /dev guest and /proc guest 1107182689 M * Bertl excellent, now let's try to start mail8n (probably best after a reboot) 1107182765 M * wfamy ok i reboot see you soon :-) 1107182795 Q * wfamy Remote host closed the connection 1107182995 M * Snow-Man vserver blah remove should probably bail if the vserver is running. 1107183013 J * wfamy bill@lns-vlq-48-mar-82-251-25-178.adsl.proxad.net 1107183028 A * micah adds echo "/usr/local/lib/util-vserver/vshelper" >/proc/sys/kernel/vshelper to /etc/init.d/vprocunhide 1107183030 M * wfamy be back 1107183043 M * Bertl Snow-Man: yes, just that ther is? no vserver blah remove yet ... 1107183059 M * Bertl wb wfamy! 1107183061 M * Snow-Man Bertl: oh, heh, well, whatever it was the guy on the list was talking about. :) 1107183085 M * Bertl hopefully your mail8 server wasn't started automagically ;) 1107183258 M * wfamy vserver mail8n start tell me there are problem with vshelper 1107183293 M * Bertl which means that the tools wherent isntalled completely/properly ... 1107183311 M * Bertl s/wherent/weren't/ 1107183320 M * wfamy arghhhhhhhh 1107183355 Q * mbucc Read error: Connection reset by peer 1107183376 M * Bertl btw, shouldn't there be debian packages for the alpha tools by now? 1107183397 M * Bertl (I mean official ones) 1107183540 M * Snow-Man You'd have to ask the maintainer, I've already started using the unofficial ones. 1107183605 M * Bertl guess that's the way to go for now .. *sigh* 1107183683 M * Snow-Man Well, personally I'd certainly like for them to be official sooner rather than later. 1107183807 Q * v00dY Remote host closed the connection 1107184235 Q * serving Ping timeout: 480 seconds 1107184602 M * micah Bertl: I was just wondering that myself 1107184611 M * micah Snow-Man: there are unofficial alpha-tools debian packages? 1107184699 M * micah Snow-Man: I am a debian developer and I might be able to help that process if I can figure out what the situation is with them 1107184733 M * Bertl it seems we have about 3-5 debian developers here, but still no official 1.9.x packages ;) 1107184772 M * Bertl (well, I doubt folks are able to find the inofficial either ;) 1107184794 M * micah it seems like there is only 0.30 -- which is stable? 1107184829 M * micah the debian developers should co-maintain these packages and setup a mailing list 1107184871 M * Bertl wfamy: are you still here? or did you give up? 1107184919 M * micah Bertl: it does seem like there is a 0.30.196 1107184925 M * wfamy still here 1107184936 M * wfamy try to find any solution 1107184948 M * Bertl okay, take your time, just was worried ;) 1107185130 M * wfamy packages.debian.org seems to be down :-( snipp 1107185167 M * micah wfamy: yes, they should be back in about 30 minutes 1107185262 J * serving ~serving@213.186.170.18 1107185336 M * Bertl wb serving! 1107185606 M * micah Bertl: it looks like there are official packages of util-vserver 0.30-196 1107185609 M * micah for debian 1107185624 M * BWare noon y'all :) 1107185667 M * Bertl 0.30-196 or 0.30.196-X ? 1107185815 M * BWare is the bz2 of 1.2.10 patch broken, or is it me ? 1107185819 M * micah Bertl: actually... I am not positive of that yet 1107185839 M * BWare patching file linux-2.4.29/net/socket.c 1107185839 M * BWare patching file linux-2.4.29/net/unix/af_unix.c 1107185839 M * BWare patch unexpectedly ends in middle of line 1107185840 M * BWare patch unexpectedly ends in middle of line 1107185840 M * BWare m 1107185847 M * Bertl BWare: url? 1107185854 M * BWare linux-vserver.org 1107185876 M * micah BWare: did you clean your source first? 1107185876 M * Bertl that's the main wiki page, no patch ;) 1107185922 M * BWare hehehe 1107185925 M * BWare http://www.13thfloor.at/vserver/s_release/v1.2.10/patch-2.4.29-vs1.2.10.diff.bz2 1107185987 M * micah Bertl: I want to create a reference vserver and then create small vserver instances that use unified elements from the reference (maybe I just want to run apache for a user, so just a vserver with an unified apache instance). 1107186009 M * micah I knew how to do this with the stable (1.2) tools, and I think I understand how to create a reference vserver with alpha 1107186033 M * Bertl BWare: $ bzcat patch-2.4.29-vs1.2.10.diff.bz2 | md5sum 1107186033 M * Bertl 35afc6fea7bbd77516aaea6395cbcb23 - 1107186053 M * micah but I dont understand how to install those specific elements in a "child" of the reference 1107186057 M * Bertl (this one is working fine here ;) 1107186107 M * Bertl micah: please elaborate! (I don't know what you are talking about atm ;) 1107186122 M * micah Bertl: sorry, I am maybe not very clear :) 1107186133 M * micah Bertl: I will explain more... 1107186194 M * micah I was using the 1.2 version before, and several months ago you helped me understand how to do things the best way 1107186199 M * micah this was: 1107186206 M * micah 1. create a reference vserver 1107186235 Q * serving Ping timeout: 480 seconds 1107186236 M * micah this reference server is not ever touched by people 1107186244 M * micah except for me to occasionally upgrade the packages 1107186272 M * micah 2. create individual vservers that use the reference server's packages for its packages 1107186298 M * micah this was done with immutable linkage inversion unification scripts that I wrote 1107186353 M * micah does this make sense so far? 1107186379 M * Bertl yep 1107186383 M * micah ok :) 1107186424 M * micah so I've finally made the leap into alpha tools and am trying to understand how I do the same with these tools 1107186437 M * BWare sorry bertl; lynx uses something totally diffent as url: Downloaded link: http://www.13thfloor.at/vserver/s_release/v1.2.10/linux-vserver-1.2.10.tar.bz2 1107186438 M * micah I see how to unify entire vservers, and individual directories 1107186462 M * BWare md5sum linux-vserver-1.2.10.tar.bz2 1107186462 M * BWare a0ec85c541557b96c5b33815eced46d2 linux-vserver-1.2.10.tar.bz2 1107186485 M * Bertl BWare: I didn't md5sum the bz2 ;) 1107186486 A * BWare needs more sleep at night 1107186525 M * Bertl but check http://www.13thfloor.at/vserver/s_release/v1.2.10/linux-vserver-1.2.10.tar.bz2.md5 1107186565 M * Bertl micah: basically you would do everything just the same (regarding unification, the changed where minimal) 1107186591 M * micah Bertl: I was not using vunify before 1107186613 M * Bertl well, you do not have to use it now ;) 1107186689 M * micah Bertl: hummm I thought that the alpha tools made that process much easier to manage 1107186712 M * Bertl yes, and vunify will probably do what you want ... 1107186841 M * micah hmm, I am confused 1107186862 M * micah you say I do everything the same, but vunify will make it easier? 1107186923 M * Bertl no, I said, that you have the choice: a) just make a small modification and use your existing framework, or b) switch to vunify and let that tool do it for you ... 1107186983 M * BWare Bertl: patch applied after tar -jxf :( 1107186990 M * BWare it was me 1107187011 M * Bertl k, everything fine then ;) 1107187016 M * BWare anyway, is ultrasparc supported by linux vserver ? 1107187035 M * micah Bertl: oh, I am setting up a new system and I want to learn the new (better) way, so I do not want to use my old framework, but use the new system 1107187036 A * BWare is missing the configuration option in make menuconfig 1107187106 M * Bertl hmm, you mean the vserver menu doesn't show up? 1107187218 M * wfamy when i configure my utils-vserver-0.30.196 it tells me Use dietlibC: no (you have been warned) 1107187225 M * wfamy is it a problem? 1107187253 M * BWare Bertl: yep 1107187274 M * BWare might have forgotten a dependency ? 1107187324 M * Bertl wfamy: well, it makes your vserver handling unreliable IIRC, but it should be fine for a test ... 1107187362 M * Bertl BWare: checking ... 1107187394 M * Bertl micah: good, then check what vunify is able to do for you ... 1107187408 M * micah wfamy: are you on debian? if so you should do apt-get intall dietlibc dietlibc-dev e2fslibs-dev iproute vlan 1107187460 M * Bertl micah: hmm, or more precisely vserver unify ... 1107187488 M * wfamy i will install it now 1107187662 M * micah Bertl: it seems like the vunifying directories is not useful, but if I set up a separate vserver for each service (such as an apache vserver, a mysql vserver, etc.) then I can unify individual vservers multipl times to get the pieces I need (for example, if I want to create a vserver that has apache and mysql then I can do a multiple unification with the apache vserver and the mysql vserver)? 1107187749 M * Bertl yes, why not ... but you could also do a default server, then create a derived apache server, mail server ... whatever from that and unify any other vserver to that ... 1107187909 M * micah Bertl: so have a master reference server, and then child references from that... how would I create those children? 1107188002 M * Bertl by simply copying them, then unifying with the master ... 1107188162 M * micah I do not understand... I think of the "master" reference server as having apache and mysql installed, then the children would just be a mysql vserver, and an apache vserver. 1107188184 M * micah is this the same as you are thinking? 1107188214 M * Bertl no, but that is fine too ... 1107188249 M * Bertl the unification will just 'unify' the files which are shared .. so it should work fine ... 1107188286 M * micah well, it is not fine, because I have no way to easily only install apache into the child reference vserver, and mysql into the child reference server through unification 1107188354 M * Bertl that is why I would start the other way round ... 1107188366 M * Bertl make a minimal 'shared' master template ... 1107188375 M * micah oooh 1107188385 M * micah I make a master skeleton 1107188392 M * Bertl create 'clones' for each specific 'child' category (e.g. http, smtp, etc ...) 1107188419 M * Bertl then install the designated stuff ... 1107188429 M * micah then unify back into the master? 1107188433 M * Bertl unify them with the master template ... 1107188451 M * Bertl and then make your 'running' copy of the child template 1107188468 M * Bertl again unify that to the master 'or' child (shouldn't really matter) 1107188650 M * Bertl does this sound reasonable? 1107188673 M * micah ok, to make sure I understand, let me re-say what you said and you tell me if I am wrong :) 1107188704 M * Bertl k, ;) 1107188708 M * micah 1. create a minimal master reference template (vserver master-reference build -m skeleton) 1107188744 M * micah 2. create clones of the master reference template (cp -a /vserver/master-reference /vserver/mysql-clone) 1107188777 M * micah 3. install service into clones (vserver mysql-clone enter; apt-get install mysql-server) 1107188779 M * Bertl (skeleton will be a little too less) 1107188801 M * micah (ok, I will maybe do a debootstrap, and then clean things out I do not need) 1107188891 M * micah 4. unify the clones with the master template (vunify --manually /vservers/clone-mysql) 1107188923 M * Bertl yep 1107188932 M * micah 5. create "actual" vservers from the clones, by making reference vservers through full vserver unification 1107188951 M * Bertl (except that it's not vunify, but vserver ... unify) 1107188986 M * micah (mkdir -p /etc/vservers/actual-mysql/apps/vunify; ln -s /etc/vservers/clone-mysql /etc/vservers/actual-mysql/apps/vunify/refserver.00) 1107188999 A * micah makes a note on #4 from what you said 1107189015 M * wfamy can't compile with ./configure --enable-dietlibc 1107189021 M * micah 6. periodically update the "master template reference vserver" and then unify the clones off of that 1107189045 M * micah 7. unify the actual-vservers from the clones 1107189048 M * Bertl wfamy: forget about dietlibc for now ... 1107189062 M * wfamy bertl: ok 1107189101 M * micah Bertl: sounds like a plan ) 1107189110 M * micah now I will attempt to do it, and then document it 1107189127 M * Bertl ok, great! 1107189205 M * micah Bertl: when I do #2 (create clones from master reference), I will need to not only do a cp -a, but also somehow get the correct information installed in /etc/vservers/clone-mysql wont I? 1107189325 M * Bertl yep, I would suggest to create the config with the skeleton method first, then copy over the files 1107189347 M * Bertl (unfortunately ther is no 'copy' build method yet) 1107189364 A * Bertl .o( hmm, well maybe there is in the new tools ;) 1107189399 M * micah :) 1107189414 M * Bertl I'm still with 0.30.196 ... 1107189446 M * micah I am ahead of you 1107189450 M * micah ;) 1107189472 M * BWare Bertl: clean vanilla kernel + patch still won't show the config menu 1107189474 M * Bertl yeah, but not for very much longer ... 1107189494 M * BWare Bertl: It won't even run a make dep :( 1107189504 M * micah BWare: it sounds like you have a tainted vanilla source 1107189520 M * Bertl BWare: hmm, where does the make dep fail? 1107189545 M * micah BWare: I just did this the other day and it worked... I would suggest that you save your .config from your kernel source, then remove your linux source directory, untar a fresh copy and start from the beginning 1107189554 M * BWare at the beginning 1107189564 M * micah I've had this problem where strange things happen with make dep, when I re-do it from a fresh untar, it works fine 1107189573 M * BWare well... the vanilla kernel does complete make dep 1107189584 M * BWare vanilla + patch won't 1107189625 M * BWare Bertl: tried that already 1107189633 M * BWare to no avail 1107189642 M * micah BWare: which patch are you using? 1107189649 M * BWare make dep 1107189649 M * BWare make -C arch/sparc64/kernel check_asm 1107189650 M * BWare make[1]: Entering directory `/usr/src/linux-2.4.29-vs1.2.10/arch/sparc64/kernel' 1107189650 M * BWare sparc64-linux-gcc -E -D__KERNEL__ -I/usr/src/linux-2.4.29-vs1.2.10/include -P tmp.c -o tmp.i 1107189651 M * BWare In file included from tmp.c:1: 1107189656 M * BWare /usr/src/linux-2.4.29-vs1.2.10/include/linux/config.h:4:28: linux/autoconf.h: No such file or directory 1107189656 M * BWare make[1]: *** [check_asm] Error 1 1107189657 M * BWare make[1]: Leaving directory `/usr/src/linux-2.4.29-vs1.2.10/arch/sparc64/kernel' 1107189657 M * BWare make: *** [check_asm] Error 2 1107189658 Q * mhepp Quit: mhepp caught signal: Autobus error 1107189664 M * Bertl ah, okay, forget it! 1107189669 M * micah ah, this is a sparc64 1107189677 M * BWare that's what I said :) 1107189681 M * Bertl 2.4/1.2 kernel does not support it 1107189684 M * micah BWare: you don't have autoconf support 1107189703 M * micah linux/autoconf.h: No such file or directory 1107189732 M * Bertl but micah is right, that is not patch related 1107189759 M * BWare weird than that the vanilla completes the make dep without problems 1107189760 M * Bertl it is most likely related to the fact that your /usr/inclöude/linux and /usr/include/asm are symlinks into /usr/src/linux 1107189798 T * services.oftc.net http://linux-vserver.org/ | latest stable 1.2.10, devel 1.9.{3,4-rc4}, ng8.13 1107189799 M * Bertl after a make distclean or mrproper those links are obsoleted 1107189832 M * BWare ls -lad /usr/include/linux 1107189832 M * BWare drwxr-xr-x 13 root root 12288 Dec 22 10:21 /usr/include/linux 1107189836 T * services.oftc.net http://linux-vserver.org/ | latest stable 1.2.10, devel 1.9.{3,4-rc4}, ng8.13 1107189838 M * BWare same for asm 1107189882 M * BWare and the vanilla does compile without errors 1107189892 M * BWare it breaks after running the patch 1107189913 M * Bertl interesting ... anyway you wont be able to get 1.2.10 up and running on a sparc ... 1107189987 M * BWare any other version that is supported ? 1107189996 M * Bertl yep, sure, 1.9.x is supported 1107190024 M * Bertl but 1.2.10 didn't have the syscall flexibility required for non x86 archs ... 1107190033 M * BWare 1.9.x is 2.6.x kernel IIRC 1107190038 M * Bertl yep 1107190056 M * BWare hmzz... i wanted to skip the 2.6 on this sparc 1107190064 M * BWare for now anyway 1107190119 M * Bertl well, you can 'hack' on the 2.4/1.2.10 for a little, basically other archs should be working, but due to the userspace incompatibilities, I didn't bother ... 1107190142 M * BWare got to get a new lab box then for 2.6 experimenting 1107190173 A * BWare ain't no kernel hacker ;) - no intentions to become one either 1107190260 J * _Plug ~plug@datadot.net 1107190291 M * Bertl welcome _Plug! 1107190314 Q * Plug Ping timeout: 480 seconds 1107190320 N * _Plug Plug 1107190433 M * wfamy reconfigure reinstall utils-0.30-196 i will try a new vserver but where can i found the new .conf file? 1107190465 M * Bertl hmm, why not continue with mail8n ? 1107190489 M * Bertl note: there is no .conf file, the new config is directory based 1107190505 M * wfamy i purge the old one 1107190627 M * Bertl okay, do as you think is best, you can find an overview of the new tools at: http://linux-vserver.org/alpha+util-vserver 1107190664 Q * _are_ Quit: Disconnecting 1107190673 M * Bertl (the Folower Page describes the new style config, but the tools will usually create them for you, see the command for mail8n creation I pasted) 1107190684 M * wfamy thx i will try this 1107190690 M * Bertl s/Folower/Flower/ 1107190715 M * Bertl okay, I'm off for now, but will probably be back in a few hours ... 1107190756 N * Bertl Bertl_oO 1107190783 M * wfamy see you 1107190931 Q * micah Quit: leaving 1107190937 J * micah micah@micha.hampshire.edu 1107191186 Q * micah Quit: 1107191307 J * micah micah@micha.hampshire.edu 1107191415 J * serving ~serving@213.186.170.250 1107191462 A * lars||| is back 1107192543 J * DuckKing ~Duck@dyn-83-157-133-238.ppp.tiscali.fr 1107192609 M * micah when I do a vserver build -m debootstrap it works fine, although I get these umount errors at the end: 1107192629 M * micah umount: /etc/vservers/.defaults/vdirbase/seedbank/dev/pts: not mounted 1107192636 M * micah same with /dev/shm and /proc/bus/usb 1107192655 M * micah is there something I should do to make that part work better, or should I just disregard them? 1107192674 Q * prae Quit: Client exiting 1107192973 Q * DuckMaster Ping timeout: 480 seconds 1107193039 M * wfamy micah: still have the vshelper problem how do you install the utils-vserver ? 1107193081 M * micah wfamy: you have a vshelper problem? 1107193082 M * lars||| micah: why don't you just use one of the premade images? 1107193103 M * micah lars|||: what pre-made images? 1107193115 M * micah wfamy: or you are saying I have a vshelper problem? 1107193179 M * wfamy i try to reinstall utils-vserver because of vshelper error but i still have the same one :-( 1107193194 M * wfamy i try to install the utils-vserver tools 1107193236 M * lars||| micah: http://www.lycos-vds.com/dists/ 1107193244 M * wfamy i run ./configure in util-vserver-0.30.196 directory then make install 1107193274 M * lars||| micah: or http://www.marlow.dk/site.php/tech/vserver 1107193307 M * wfamy it seem to be ok but still sending me the error. So do you use any tricks to install the utils? 1107193316 M * micah wfamy: can you tell me the error? 1107193369 M * micah lars|||: these look like entire distribution images... I don't want that 1107193412 M * Doener micah: you can disregard those messages at the end of the debootstrap, nothing to worry about 1107193431 M * lars||| micah: ok, sorry... I just saw you got umount errors 1107193465 M * micah Doener: thanks :) 1107193485 M * micah lars|||: it looks like these images are built the same way I am doing it anyways :) but thanks for the pointer 1107193495 M * micah wfamy: I think I know the solution, but I need to know the error 1107193598 M * wfamy i do not understand i search in th eold directory. 1107193623 M * wfamy The configured vshelper '/opt/vserver//lib/util-vserver/vshelper echo /opt/vserver//lib/util-vserver/vshelper' does not match the 'vshelper' 1107193643 M * Doener wfamy: together with that message you get a bunch of solutions... 1107193687 M * Doener and from what i can tell, i'd say you double-pasted one of them... 1107193697 M * micah wfamy: I added the echo line to my vprocunhide startup script 1107193715 M * DaCa micah: I think the umount problem is in debootstrap itself, you get the same errors when running debootstrap only, I added a || true on line 567 in /usr/lib/debootstrap/functions to make newvserver work again (I am using vserver-debiantools) 1107193733 M * Doener micah: why not /etc/sysctl.conf? seems more 'clean' to me 1107193748 M * wfamy ok 1107193763 M * micah Doener: it probably is, I should do that 1107193794 M * Doener DaCa: yep, i guess the debootstrap process also tries to mount those things earlier, but the message just fly over the screen, so that one does not notice it. anyway, it causes no trouble :) 1107193806 M * wfamy micah: after echo this time it seem to work. I will try to install my daemontools and tell you 1107194127 M * lars||| what is the progress of the patch for 2.6.10 ? 1107195355 M * lars||| can somebody tell me which patch to use for 2.6.10 ? 1107195426 M * daniel_hozac 1.9.4-rc4 is probably recommended. 1107195437 M * lars||| ok 1107195548 M * micah lars|||: I ended up using 2.6.11-rc2 1107195697 M * lars||| micah: did rc4 not work? 1107195720 M * lars||| a 2.6.11 patch for 2.6.10 ? 1107195922 M * Doener 1.9.4 patches are available for 2.6.{9,10,11-rc2} 1107195928 N * Doener Doener|gone 1107196045 N * Bertl_oO Bertl 1107196045 M * lars||| micah: ok, so you use 2.6.11 with the 2.6.11-rc2 patch 1107196071 M * Bertl evening folks! 1107196100 M * lars||| welcome back Bertl! 1107196135 M * Bertl tx! 1107196206 M * Bertl wfamy: still trouble with installing the tools? 1107196247 M * Bertl rs: what are your plans regarding ngnet? 1107196252 M * lars||| so is there any known trouble with 1.9.4-rc4 and 2.6.10 or is it only micah that has a weird machine :) 1107196338 M * Bertl well, I don't know what troubles micah encountered, if any ... but 1.9.4-rc4 is supposed to work ... (the only thing missing there is a fix regarding SPINLOCK debugging and VSERVER debuggin on UP) 1107196370 M * Bertl lars|||: if you have any troubles with that, please let me know, I'll fix it in 1.9.4-rc5 ;) 1107196384 M * lars||| I will try then 1107196762 J * prae ~prae@sherpadown.net 1107196797 M * Bertl welcome prae! 1107196916 M * prae hi Bertl 1107197609 M * rs Bertl: I'll continue to work on it tomorow 1107197770 M * Bertl just wanted to know, thanks for the info! 1107197882 M * Pazzo Bertl: (little feedback) tested 2.6.11rc2-vs1.9.4-rc4 during the weekend, it's now up and running on the productional host - looks good, I'll let you know if something goes wrong ;-) 1107197913 M * Pazzo s/the productional/a productional/ 1107198246 Q * Hollow Remote host closed the connection 1107198288 M * wfamy bertl: i install them and this time the echo "..." >/proc.. seems working 1107198334 J * Hollow ~bene@home.xnull.de 1107198487 M * Bertl Pazzo: thanks for the info! 1107198504 M * Bertl wfamy: good! 1107198556 J * chrish01 ~chrish01@69.90.131.25 1107198816 Q * Hollow Quit: 1107198946 Q * rs Quit: leaving 1107198975 M * Bertl wfamy: so let's attack the 'original' issue if possible ... 1107199028 M * chrish01 hi Bertl, good weekend? 1107199040 M * wfamy bertl: I work in installing a daemontools 1107199051 M * Bertl chrish01: excellent weekend, and you? 1107199067 M * chrish01 decent =P got most of a graphical diff viewer written in C#/Gtk# 1107199079 M * Bertl wfamy: good! 1107199301 J * Hollow ~bene@home.xnull.de 1107199320 M * Bertl chrish01: new ngnet with new features is out ;) 1107199439 M * chrish01 Bertl, i caught that :) 1107199447 M * chrish01 what new features if i may ask? 1107199463 M * Bertl tcp is now working *G* 1107199469 M * chrish01 haha 1107199480 M * Bertl well, actually it's more a bugfix release than a new feature release 1107199489 M * chrish01 hey, if it works :) 1107199523 M * Bertl except for two known issues, it should work now ... completely 1107199544 M * chrish01 awesome Bertl. you rock 1107199564 M * Bertl issues are: arp isn't done for guests, and inter guest connectivity isn't implemented 1107199604 M * chrish01 ok, for arp in guests, we just need something to pass arp from outer to inner (something like proxyarp)? 1107199668 M * chrish01 garrr, no delta to 13 :) 1107199843 M * Bertl hold on, I upload a delta 1107199850 M * chrish01 too late :) 1107199855 M * Bertl k 1107199908 M * chrish01 the new vmware is pretty decent :) (5.0 rc) 1107199912 M * Bertl arp in guests is actually solved easier .. you just assign the ip on the host too 1107199926 M * chrish01 interesting 1107200022 M * wfamy bertl: when i do vserver mail10 stop i had "vkill: vc_ctx_kill(): No such process" 1107200040 M * wfamy and daemontools not yet installed 1107200132 M * Bertl well, that could mean that there just is no process 1107200137 M * Bertl which would be fine then ... 1107200184 J * hansi33 ~hansi33@193.80.53.173 1107200204 M * Bertl welcome hansi33! 1107200216 M * hansi33 hello Bertl 1107200248 M * Bertl a visit from austria, nice ;) 1107200251 M * hansi33 i make my first experience with sched_hard, but it ist not as i expected 1107200314 M * Bertl well, let's hear what happens, and what you expected then ... 1107200353 M * hansi33 i have to vservers running: one with parameters 500/90/100/0/500, the other with 500/45/100/0/500 1107200395 M * hansi33 if i start a long job on both they duritaion is exactly the same 1107200406 M * hansi33 to -> 2 1107200437 M * Bertl short question: what kernel (kernel patches) and what tools? 1107200461 M * hansi33 1.9.3 / 0.30.196 1107200499 M * Bertl okay, you enabled the hard scheduler at compile time, and added the appropriate flag to the vserver's config? 1107200520 M * hansi33 yes 1107200558 M * Bertl what does the 'job' do? 1107200581 M * hansi33 int i,j; 1107200582 M * hansi33 unsigned long long v1,v2,v3; 1107200584 M * hansi33 for (j=1;j<1000000;j++) { 1107200585 M * hansi33 v1 = 42; 1107200587 M * hansi33 v2 = v3 = -1; 1107200588 M * hansi33 for (i=0; i<1000; i++) { 1107200590 M * hansi33 v1 += i; 1107200591 M * hansi33 v2 = v1*i + v3; 1107200593 M * hansi33 v3 = v1 ^ v2; 1107200594 M * hansi33 } 1107200596 M * hansi33 } 1107200597 M * hansi33 return 0; 1107200599 M * hansi33 duration approx 20 secs 1107200611 M * Bertl hehe, looks very much like my cpuhog() ;) 1107200617 M * hansi33 yes it is 1107200630 M * Bertl okay, that should be cpubound, that is fine ... 1107200728 M * Bertl couldso what are the values you set standing for? 1107200730 M * micah hmm, I can't seem to start my vserver, when I do "vserver seedbank start" I just get a prompt back, but nothing else 1107200745 M * Bertl hansi33: s/couldso/so/ 1107200765 M * Bertl micah: does vserver-stat show it running afterwards? 1107200773 M * micah Bertl: no, it says it is stopped 1107200797 M * hansi33 what do you mean with standing for? 1107200812 J * mbucc ~konversat@host-216-153-147-194.spr.choiceone.net 1107200813 M * Bertl hansi33: maybe the output of cat /proc/virtual//sched could sched some light on this ... 1107200823 M * Bertl wb mbucc! 1107200835 M * mbucc hi 1107200839 M * hansi33 Token: 500 1107200840 M * hansi33 FillRate: 90 1107200842 M * hansi33 Interval: 100 1107200844 M * hansi33 TokensMin: 0 1107200844 M * Bertl micah: that means that your vserver didn't start any services which lasted longer than the startup 1107200845 M * hansi33 TokensMax: 500 1107200847 M * hansi33 PrioBias: 0 1107200848 M * hansi33 cpu 0: 22329 826 380 1107200850 M * hansi33 Token: 500 1107200852 M * hansi33 FillRate: 90 1107200853 M * hansi33 Interval: 100 1107200855 M * hansi33 TokensMin: 0 1107200856 M * hansi33 TokensMax: 500 1107200858 M * hansi33 PrioBias: 0 1107200859 M * hansi33 cpu 0: 22436 856 1000 1107200867 M * hansi33 oops i see, they are all 90 not 45 1107200871 M * Bertl ;) 1107200881 M * micah Bertl: I did remove as many packages as I could, so maybe I removed something important? 1107200904 M * Bertl micah: probably no service is left, make sure that syslog for example is there .. 1107200916 M * Bertl (but I guess for a template it's fine if it doesn't start either ..) 1107200930 M * micah Bertl: I removed syslog :) 1107200940 M * wfamy bertl: when i start my server with the fakeinit it do not lunch the init and daemontools does not work :-( 1107200950 M * micah Bertl: i dont really need to start the vserver, but I do need to enter it so I can do things within it 1107200965 M * micah Bertl: and if I try to enter: 'vserver ... suexec' is support for running vservers only; aborting... 1107200968 M * Bertl wfamy: very likely you need to change the 'initstyle' 1107200996 M * Bertl micah: I.c. so syslog it is ;) 1107201078 M * micah Bertl: and I can only enter a vserver if I have a running process of some kind? 1107201280 M * Bertl well, basically you could do that without, but for some reason the tools do not support that atm ... 1107201288 M * wfamy bertl: cat /usr/local/etc/vservers/mail10/flags ->fakeinit and found no option in vserver for initstyle. 1107201344 M * micah Bertl: hmm, I wonder what I can do 1107201366 M * Bertl wfamy: * /etc/vservers/vserver-name/apps/init , # 1107201368 M * Bertl # style 1107201374 M * Bertl # 1107201374 M * Bertl Contains the init-style. 1107201401 M * Bertl micah: as I said, easiest is to add _and_ enable syslogd ... 1107201469 Q * mbucc Remote host closed the connection 1107201478 J * mbucc ~konversat@host-216-153-147-194.spr.choiceone.net 1107201557 M * micah Bertl: yes, but I have no way of adding syslog 1107201576 M * micah Bertl: because the vserver is stopped, and I cannot start it, nor can I do vserver apt-get 1107201620 M * Bertl huh? 1107201632 M * Bertl the vserver pkg install .... 1107201643 M * Bertl should work in this case too (see vserver --help) 1107201667 M * Bertl (probably same for apt-get, rpm and whatever) 1107201685 M * hansi33 Bertl: I did dynamic changing of scheduler-parameters with: vsched --ctx 31 --fill-rate 45, I got a segmentation fault, but the parameters do change, its not a problem for me, but do you have a hint whats the reason for this. 1107201706 M * Bertl bug in the tools, require some argument ... 1107201728 M * Bertl (is already fixed in recent tools, but you can workaround by adding -- true ) 1107201786 M * Bertl 'vsched --ctx 31 --fill-rate 45 -- true' should work 1107201797 M * hansi33 yes it worked. 1107201825 M * micah Bertl: when I do vserver apt-get install sysklogd it simply returns the command prompt, and does not install it... the pkg install method tells me: 'vserver ... suexec' is support for running vservers only; aborting... 1107201862 M * Bertl micah: what tool version? 1107201871 M * micah Bertl: the one you don't have yet :) 1107201880 M * micah util-vserver-0.30.201 1107201886 M * Bertl ah, okay, so I'd say, bug report! ;) 1107201895 M * Bertl btw, I already have it ;) 1107201901 M * micah hehe, ok 1107201927 M * micah so maybe I should install the 0.30.196 and do the pkg install? 1107201931 M * micah to get out of this 1107201959 M * Bertl don't know if that will work, I do not thing it's a regression ... 1107202000 M * micah hmm, I remember now I could not ./configure .196 because my dietlib is 0.27-7 and it thinks it is too old 1107202133 M * wfamy bertl: th eflower page do not help me /etc/vservers/vserver-name/apps/init i do not understand what is it. 1107202175 M * micah Bertl: I solved it 1107202192 M * hansi33 is there a better configuration for schedule to don't waste cpu-time, if i use fill-rate 50 / interval 100 and there is nothing else to do on the other vservers i waste cpu-time 1107202200 M * micah Bertl: i simply did a chroot /vserver/ then installed the sysklogd package :) 1107202249 M * Bertl hansi33: that is exacltly the difference between hard and soft scheduler 1107202268 M * Bertl hard means hard, soft just adjusts the priorities ... 1107202291 M * Bertl micah: excellent! 1107202334 M * hansi33 how do i soft scheduling? 1107202354 M * Bertl use the SCHED_PRIO instead of the SCHED_HARD flag 1107202400 M * hansi33 where do i put the parameters for sched_prio? 1107202438 M * Bertl it's the same scheduler .. so your values are already there, just without sched_hard it will not be enforced just priortized 1107202463 M * hansi33 ok, thanks, i will try it now. 1107202470 M * Bertl i.e. tasks from contexts with more tokens will get higher priority 1107203029 M * micah Bertl: is it possible to "chain" unification? 1107203063 M * micah I ask because to make my mysql clone, I am required (because of package dependencies) to have a MTA installed 1107203119 M * Bertl what do you mean by 'chain'? 1107203120 M * micah so, can I make a postfix clone vserver from the reference, then the mysql clone uses the main template, as well as the postfix vserver as its unification base 1107203171 M * Bertl you can even do a mesh of unifications, if you like to ;) 1107203194 M * micah the problem is my ignorance, I dont *know* if I want to ;) 1107203199 M * Bertl (unification is nothing else than hardlinks atm) 1107203233 M * Bertl a <-> b and b <-> c (<-> is unification) 1107203254 M * Bertl means a <-> c too 1107203348 M * Bertl (where a, b and c are files) 1107203372 A * micah thinks 1107203490 A * mbucc perks up at unification talk ... 1107203496 M * micah if I have /usr/sbin/mysql in vserver a, and /usr/sbin/sendmail in vserver b and I unify them, will both vservers have /usr/sbin/sendmail and /usr/sbin/mysql when finished? 1107203697 M * micah or is the idea to make a default template master server that is not unified to the clones? 1107203705 M * hansi33 are the parameters in schedule realy the same for sched_prio? i always get the same execution time for both jobs in the vservers, the number of tokens is always 0, also if i setup fillrate 200 / interval 100 for one vserver 1107203757 M * micah unify just unifies those files that are shared, not adding new things, right? 1107203787 M * Bertl micah: yep 1107203848 M * Bertl hansi33: what machine is it running on? 1107203890 M * hansi33 p4, 2.4 GHz, gentoo linux 1107203934 M * mbucc micah: so if you have multipe clone types, it's kind of like calculating a least-common-denominator for dependencies. 1107203940 M * mbucc micah: right? 1107203948 M * mbucc micah: mysql + mta 1107203960 M * mbucc micah: mysql + mta + apache + php 1107203969 M * Bertl hansi33: HT (SMT)? 1107203970 M * mbucc etc. 1107203977 M * hansi33 no 1107204071 M * Bertl okay, let's check cat /proc/virtual//status of the context in question 1107204086 A * mbucc realizes i need to read more before typing ... 1107204107 M * hansi33 UseCnt: 14 1107204108 M * hansi33 RefCnt: 12 1107204110 M * hansi33 Flags: 0000000202000210 1107204111 M * hansi33 BCaps: ffffffffd44c04ff 1107204113 M * hansi33 CCaps: 0000000000000101 1107204115 M * hansi33 Ticks: 0 1107204116 M * hansi33 UseCnt: 14 1107204118 M * hansi33 RefCnt: 12 1107204119 M * hansi33 Flags: 0000000202000210 1107204121 M * hansi33 BCaps: ffffffffd44c04ff 1107204122 M * hansi33 CCaps: 0000000000000101 1107204124 M * hansi33 Ticks: 0 1107204145 M * Bertl hmm, that looks good, actually ... 1107204160 N * chrish01 chris|lunch 1107204169 M * Bertl and your scheduler config is different now? 1107204177 M * hansi33 is there a problem with ht? because my next machine for production use will have it? 1107204189 M * Bertl (i.e. it's different between the two vservers ?) 1107204221 M * hansi33 Token: 0 1107204222 M * hansi33 FillRate: 200 1107204222 M * Bertl hansi33: no, just on a HT/SMP machine two threads will always run at 100% 1107204224 M * hansi33 Interval: 100 1107204225 M * hansi33 TokensMin: 0 1107204227 M * hansi33 TokensMax: 500 1107204228 M * hansi33 PrioBias: 0 1107204230 M * hansi33 cpu 0: 22597 842 0 1107204231 M * hansi33 Token: 0 1107204233 M * hansi33 FillRate: 90 1107204235 M * hansi33 Interval: 100 1107204237 M * hansi33 TokensMin: 0 1107204238 M * hansi33 TokensMax: 500 1107204240 M * hansi33 PrioBias: 0 1107204241 M * hansi33 cpu 0: 43622 821 0 1107204252 M * Bertl that looks like a bug ... 1107204285 M * micah mbucc: I am not sure actually 1107204303 M * Bertl hansi33: sec, trying to reproduce here ... 1107204490 M * micah Bertl: if I create a simple master template vserver, then clone that into a vserver called mysql-clone. I then enter mysql-clone and install mysql. It seems pointless to unify mysql-clone with the master template 1107204671 A * micah tries to get his brain to work right 1107204698 M * Bertl yes, it's pointless, but it won't hurt, that's the point ;) 1107204751 M * micah ah, ok :) 1107204782 M * daniel_hozac how can it be pointless and have a point at the same time? 1107204783 M * daniel_hozac ;) 1107204788 J * yarihm ~yarihm@80-218-7-247.dclient.hispeed.ch 1107204796 M * Bertl welcome yarihm! 1107204812 M * micah ok, so when I try to create the mysql-clone, I can't install mysql because it depends on a MTA... can I somehow get that MTA reference into this clone? 1107204835 Q * mbucc Ping timeout: 480 seconds 1107204848 J * mbucc ~konversat@host-216-153-147-194.spr.choiceone.net 1107204873 M * Bertl hansi33: yeah, I can confirm, it is a kernel bug, the token recalculation is optimized away if sched hard is disabled ... 1107204938 M * micah or, maybe it is better for my master_simple template to have a MTA installed 1107204941 M * hansi33 is there another way to give diffrent priorities to diffrent vservers 1107204990 M * Bertl as it is a bug, it will be fixed very soon, atm, I guess using the hard scheduler is the best choice ... 1107205004 M * wfamy i create mail13 with "vserver mail13 build -m debootstrap --hostname mail13 --interface eth1:192.168.0.113/24 --context 42 --flags fakeinit -- -d sarge" 1107205043 M * wfamy when i start if it i have no trace in ifconfig may have eth1:mail13 1107205068 M * Bertl which is expected ... 1107205083 M * Bertl the new tools do not use aliases by default, if you want that you have to use: 1107205099 M * Bertl --interface mail13=eth1:192.168.0.113/24 1107205125 M * Bertl wfamy: ip addr ls will list the ips anyway 1107205199 M * wfamy it is because mail13 network do not work 1107205450 M * wfamy mail13:/etc# ping 192.168.0.1 connect: Invalid argument. 1107205458 M * wfamy i going slighty mad ! 1107205493 M * Bertl don't use ping for such checks, and if, specify the ip with -I 1107205507 M * Bertl ping -I 192.168.0.113 192.168.0.1 1107205532 M * wfamy ping -I 192.168.0.113 192.168.0.1 bind: Cannot assign requested address 1107205558 M * Bertl so 192.168.0.113 is not available in that vserver ... 1107205571 M * Bertl btw, you are not using --context 42 for all of your vservers, right? 1107205587 M * Bertl (--context is supposed to be unique) 1107205636 M * wfamy ok 1107205825 M * mugwump http://isec.pl/vulnerabilities/isec-0021-uselib.txt # anyone checked if this allows vserver users to break out into master ctx? 1107205868 M * Bertl nope, but a) it's already fixed, and b) I don't think so ... 1107205934 M * wfamy i edit the mail13/context with 420 then vserver mail13 start and enter them same problem 1107205974 M * Bertl check that the ip is assigned on the hsot (with ip addr ls) 1107205990 M * Bertl and verify the grep ipv4 /proc/self/status shows the ip too 1107206110 Q * serving Ping timeout: 480 seconds 1107206141 M * Bertl hansi33: if you want to test, I have a preliminary fix ready ... 1107206167 M * wfamy doug:/usr/local/etc/vservers/mail13/interfaces/0# for i in `ls`;do echo $i;cat $i;done dev eth1 **ip 192.168.0.113 ** prefix 24 1107206173 M * hansi33 yes, i would like to test it 1107206193 M * wfamy doug:/usr/local/etc/vservers/mail13/interfaces/0# grep ipv4 /proc/self/status 1107206193 M * wfamy ipv4root: 0 1107206193 M * wfamy ipv4root_bcast: 0 1107206318 M * Bertl hansi33: http://vserver.13thfloor.at/Experimental/RC-1.9.4/delta-schedprio-fix01.diff 1107206333 M * Bertl it's against 1.9.4-rc4 but should apply to older ones too 1107206401 M * Bertl wfamy: that was on the host, right? 1107206430 M * Bertl try from inside the guest (the grep) 1107206473 M * wfamy mail13:/# grep ipv4 /proc/self/status 1107206473 M * wfamy ipv4root: 7100a8c0/00ffffff 1107206473 M * wfamy ipv4root_bcast: ffffffff 1107206530 M * Bertl hmm, and ping -I 192.168.0.113 192.168.0.1 doesn't work inside that vserver? 1107206548 M * wfamy mail13:/# ping -I 192.168.0.113 192.168.0.1 1107206548 M * wfamy bind: Cannot assign requested address 1107206557 M * Bertl (ah, you probably are missing a flag .. sec) 1107206603 M * Bertl more precisely a ccap 1107206609 M * Bertl DECL("icmp", VC_VXC_RAW_ICMP), 1107206635 M * Bertl * /etc/vservers/vserver-name/# 1107206636 M * Bertl # ccapabilities 1107206648 M * Bertl add 'icmp' there ... 1107206773 M * Bertl hansi33: you have to adjust the scheduler settings slightly, i.e. so that the values do not overbook the scheduler (as the context are not paused when the tb is empty, they will be refilled at the specified rate) 1107206851 M * hansi33 ok, my machine is compiling the kernel. i will try diffrent settings. 1107206905 M * wfamy doug:/usr/local/etc/vservers/mail13# cat ccapabilities 1107206906 M * wfamy icmp 1107206924 M * wfamy doug:/usr/local/etc/vservers/mail13# vserver mail13 start 1107206924 M * wfamy /usr/local/sbin/vserver: line 543: ip: command not found 1107206924 M * wfamy /usr/local/sbin/vserver: line 545: ip: command not found 1107206924 M * wfamy Unknown ccap 'icmp' 1107206958 M * Bertl hmm, hmm, ip: command not found, guess you forgot to install iproute2 1107206960 J * Rusty` ~root@3e70dd23.adsl.enternet.hu 1107206964 M * Rusty` hi guys 1107206968 M * Bertl welcome Rusty`! 1107207032 Q * Rusty` Quit: 1107207076 M * wfamy apt-get install iproute 1107207087 M * wfamy doug:/usr/local/etc/vservers/mail13# vserver mail13 start 1107207087 M * wfamy Unknown ccap 'icmp' 1107207128 M * Bertl check in your tools (and look at http://savannah.nongnu.org/cgi-bin/viewcvs/util-vserver/util-vserver/lib/ccaps-v13.c?rev=HEAD) 1107207147 M * Bertl if your tools do not define the ccap, you can fall back to the numeric value ... 1107207270 M * Bertl hansi33: not convinced that the patch is 100% what we want ... 1107207287 M * hansi33 why? 1107207304 M * Bertl because I probably missed something, checking now ;) 1107207406 J * matta-laptop ~matta@69.93.28.254 1107207419 M * Bertl could it be? matta? 1107207422 M * matta-laptop hello 1107207428 M * matta-laptop how are you doing? 1107207435 M * Bertl you got a laptop? fine, thanks! 1107207447 M * matta-laptop oh, heh 1107207453 M * matta-laptop sorry, i use this for work also 1107207483 M * matta-laptop quick question 1107207500 M * matta-laptop i know you have the vserver stuff ported to UML arch to help kernel debugging 1107207510 M * matta-laptop has anyone ported it to the xen arch for the same thing? 1107207553 M * matta-laptop the code is pretty much arch independent right, it's just adding a few definitions, etc right? 1107207567 M * Bertl not that I know of ... but I guess xen will not be so exotic that it would not work ... 1107207615 M * matta-laptop hr 1107207707 M * matta-laptop why the 3GB lowmem patch added to vserver? it's required? 1107207727 M * Bertl no but it is very usefule for 1GB or 2G systems 1107207730 M * Bertl +B 1107207740 M * matta-laptop oh 1107207749 M * matta-laptop might want to make that a separate patch :) 1107207768 M * Bertl you can turn off highmem with that, and it is a separate patch ;) 1107207773 M * matta-laptop it is? 1107207792 M * matta-laptop I downloaded the patch-2.6.9-vs1.9.3.diff file and it's in there 1107207801 M * Bertl yeah, it was also submitted on lkml ... 1107207812 M * Bertl (and we are at 1.9.4-rc4 now ;) 1107207823 M * matta-laptop oh, ok 1107207826 M * matta-laptop didn't see that in the changelog 1107207840 M * Bertl see topic ;) 1107207887 M * wfamy bertl: rgrep ICPM * 1107207890 M * wfamy ccaps-v13.c: DECL("ping", VC_VXC_ICMP_PING), 1107207899 M * matta-laptop ok 1107207900 M * matta-laptop yeah 1107207907 M * matta-laptop the i386 specific code seems minimal 1107207909 M * wfamy adding ping to ccap and vserver mail13 start work 1107207910 M * matta-laptop i'll give this a shot 1107207913 M * Bertl wfamy: okay, then let's use that one ;) 1107207937 M * Bertl matta-laptop: that is why it supports about 8 different platforms ;) 1107207980 M * matta-laptop i'll get back to you if it works 1107207986 M * matta-laptop well, i guess i'll idle here 1107207990 M * matta-laptop need to stay on for work anyhow 1107207999 M * Bertl make it so, you are welcome! 1107208042 M * wfamy starting qmail install with daemontools. 1107208213 J * _are_ foobar@pD95EE94B.dip.t-dialin.net 1107208221 M * Bertl welcome _are_! 1107208227 M * _are_ hi 1107208269 M * wfamy hi 1107208296 M * _are_ found no quick way to enable rpc.lockd (or something similar) within a vserver. the binary just terminates again. 1107208337 M * matta-laptop ok, the patch applies over a xen tree with no problems 1107208346 M * _are_ I think I might just miss a CAP, but I somehow lost where I can see what CAPs actually are in a vserver, I think I have seen some /proc entry 1107208370 M * Bertl grep Cap /proc/self/status 1107208402 M * _are_ hehe, i looked for something containing cap or vserver in name. :-> 1107208423 M * Bertl (you execute it _inside_ the vserver) 1107208430 M * _are_ yes, ofc 1107208471 M * micah mbucc: I am wondering if you understand the multiple clone unification idea? 1107208659 M * chris|lunch Bertl, xfrm_policy.o is broken in ngnet 8.13 1107208687 M * Bertl hmm, how does it come that you use that? 1107208704 M * Bertl (and yes, I know) 1107208715 M * chris|lunch one sec, let me track dependency 1107208752 N * chris|lunch chrish01 1107208792 M * chrish01 hmm, is that ipsec stuff? 1107208816 M * Bertl hansi33: for whatever reason the tb isn't refilled here, what about your setup, does it work there? I guess no ... 1107208824 M * chrish01 and yes i need it since the vservers will be using IPSec KAME tools 1107208847 M * Bertl hmm, okay, could you upload the error message please? 1107208851 M * chrish01 sure 1107208878 M * hansi33 Bertl: I am lucky with your patch, my configuration for fillrate/interval is 90/100 and 10/100, if i run on both execution time is 23 to 46 seconds 1107208890 M * chrish01 Bertl, http://pastebin.com/235899 1107208902 M * hansi33 if i run it alone execution time is 21 seconds on both vservers 1107208912 M * Bertl hansi33: hmm, well, it is supposed to work, but it fails on my test setup :/ 1107208948 M * Bertl chrish01: tx 1107208954 M * hansi33 is there anything i can do for you? 1107208965 M * chrish01 np 1107208992 M * Bertl hansi33: yes, please check the */sched stuff are there tokens in any of the contexts while the hog runs? 1107209063 M * Bertl chrish01: okay, just remove the entire vxdprintk (lines 726-728) for now 1107209086 M * hansi33 yes they are in order, the slow vserver goes donw to 0, the fast one is about between 300 and 400 1107209090 M * chrish01 ok 1107209119 M * Bertl hansi33: okay, so that is fine than, remains the question: what am I testing right now?! 1107209182 M * wfamy bertl: when instaling daemontools -- package/run.inittab: line 11: kill: (1) - No such process 1107209195 M * wfamy ps aux do not show them :-( 1107209212 M * Bertl wfamy: ah, interesting! 1107209230 M * Bertl okay, the server you created, how large is it? 1107209293 M * micah Bertl: I am afraid creating clones sets me up for a inter-clone dependency hell 1107209307 M * wfamy cd /; du -h 1107209310 M * wfamy 1.2G 1107209327 M * wfamy du -h /proc 897M /proc/ 1107209335 M * Bertl try du -skx /path/to/vserver 1107209349 M * Bertl (on the host) 1107209369 M * wfamy doug:/vservers# du -skx mail13/ 1107209369 M * wfamy 273916 mail13/ 1107209395 M * Bertl okay, is there any sensitive information in that guest? 1107209421 M * wfamy no i just install it 1107209455 M * Bertl good, do you ahve some place to upload a compressed version (will be about 80-100MB in size) 1107209476 M * wfamy sure 1107209488 M * Bertl then do the following please: 1107209518 M * Bertl dump 0zf /somewhere/mail13.dump /path/to/mail13 1107209518 M * micah Bertl: the reason I say this is because ... lets say in the end I wanted to create a vserver for a group that wants a drupal installation... in order to do this, I need to have the following clone unification dependencies: http://pastebin.ca/5051 1107209705 M * wfamy mail13.dump = 150Meg 1107209728 M * Bertl well, okay, for me, if it is fine for you ... 1107209824 M * mbucc micah: i am reading up on unification right now ... 1107209906 M * wfamy do you whant me end you the dump? 1107209914 M * wfamy send you 1107209933 M * Bertl well, if you can upload it somewhere (I'll grab it via http or ftp) 1107209946 M * mbucc micah: http://archives.linux-vserver.org/200202/0211.html 1107209996 M * Bertl micah: looks good ... 1107210051 M * wfamy ok i will do it tomorow; now go to bed. 1107210062 M * Bertl k, have a good night! 1107210072 M * wfamy bye 1107210076 Q * wfamy Quit: Leaving 1107210107 M * micah Bertl: it seems problematic, because then I need to set up unification for each of those clones 1107210139 M * micah when it seems easier to create a "base reference server" that has apache, php, mysql, mta, perl (since 90% of vservers will use those) 1107210231 M * micah then create client vserver installations off of that base reference 1107210249 M * micah unless there is a significant advantage to having the individual service clones, I only see some administrative overhead 1107210297 M * micah (it is very possible I do not understand something here though!) 1107210361 M * Bertl well, unifying to the drupal-clone will automatically unify to all the others, if the drupal clone was unified to them ... 1107210411 M * micah Bertl: yes, but the drupal clone will have to unify to the php4, and the php4 will have to unify with apache, and ... 1107210422 M * Bertl micah: but sure, you can have a single base reference server too ... why not? you came up with the specialized templates idea after all ;) 1107210448 M * micah it was me? :o 1107210479 M * Bertl 21:23 < micah> Bertl: is it possible to "chain" unification? 1107210503 M * micah ah, yes, but that question was because I thought through what you were suggesting 1107210521 M * micah 1107188366 M * Bertl make a minimal 'shared' master template ... 1107210528 M * micah 1107188392 M * Bertl create 'clones' for each specific 'child' category (e.g. http, smtp, etc ...) 1107210531 M * micah etc. 1107210534 M * micah 1107188419 M * Bertl then install the designated stuff .. 1107210692 M * Bertl 1107187909 M * micah Bertl: so have a master reference server, and then child references from that... how would I create those children? 1107210712 M * micah whose brain child was this! 1107210713 M * Bertl 1107188002 M * Bertl by simply copying them, then unifying with the master ... 1107210713 M * micah :) 1107210716 M * Bertl 1107188162 M * micah I do not understand... I think of the "master" reference server as having apache and mysql installed, then the children would just be a mysql vserver, and an apache vserver. 1107210732 M * Bertl so, before we start making the irc log recursive ;) 1107210743 M * Bertl let's go on and do something useful ;) 1107210746 M * micah heheheh 1107211230 Q * wishes Quit: Changing server 1107211595 M * Bertl hansi33: do you have CONFIG_VSERVER_HARDCPU_IDLE=y ? 1107211608 M * hansi33 no 1107211886 M * Bertl well, okay, I guess I have to give up for today .. here the processes behave exactly the same as before ... 1107211896 M * Bertl (before the patch, that is :/) 1107211914 M * Bertl hansi33: final question for today, what kernel patch do you have? 1107211971 M * hansi33 1.9.3 1107211979 M * Bertl okay, thanks ... 1107211991 M * Bertl have a good night everyone, cya tomorrow ... 1107212015 N * Bertl Bertl_zZ 1107212100 M * chrish01 night Bertl_zZ, thanks again :) 1107212396 Q * mbucc Ping timeout: 480 seconds 1107212848 J * serving ~serving@213.186.180.153 1107213823 Q * yarihm Quit: Leaving 1107214298 J * serving_ ~serving@213.186.180.153 1107214298 Q * serving Read error: Connection reset by peer 1107215253 Q * prae Quit: leaving 1107215276 Q * lars||| Remote host closed the connection