1115424305 M * albeiro hm, could you guys check two things for me ? namely, tune2fs option to enable writeback journal and tree indexing ? ;] 1115424316 M * albeiro there is no man on instalation cd, shame 1115424416 M * Bertl albeiro: use the source, or man ;) 1115424424 M * Doener -o journal_data_writeback -O dir_index 1115424441 M * albeiro no man. i almost used the source ;;] 1115424443 M * albeiro thx 1115424454 M * albeiro i hope this options are a good one ;] 1115424460 M * Doener you should have elinks or at least lynx... use google ;) 1115424492 M * albeiro google shows only outdated manuals 1115424525 Q * atsab Read error: Operation timed out 1115424537 J * atsab ~as@lotes.vtu.lt 1115424581 M * Doener hmm... first hit i get looks good... 1115425696 M * romke can I use normal quota tools inside vserver? 1115425744 M * Bertl yup, that works ... 1115425759 M * Bertl but you need to do some adjustments to make it secure 1115425860 M * romke Bertl: I need it read only (is there any howto?) 1115425928 J * eran ~eran@ip-213-49-106-244.dsl.scarlet.be 1115425942 M * Bertl welcome eran! 1115425946 M * Bertl romke: read only? 1115426032 M * romke Bertl: I only need read how much chosen uid have curently alocated 1115426040 N * eran terr 1115426045 M * terr ;-) 1115426066 M * terr Hello. 1115426107 M * Doener hi terr 1115426121 A * Doener makes a note eran==terr==Gilles 1115426132 M * terr Yes, right. 1115426133 M * Doener ;) 1115426172 M * terr I should learn how to use IRC. 1115426181 M * romke Bertl: kernel does all the math when I put any file in guest - right? (math=checking for limits, updating usage) 1115426227 M * terr I'm trying another IRC client... 1115426268 M * romke Bertl: as far as I tested limit set on host is preserved on guest, but I'd like that user could check his/her quota usage by executing quota prog 1115426275 M * Bertl romke: yes, ekrnel does the math, and probably you could even make it readonly with some tricks ... 1115426318 M * Bertl but by default there are files containing the quota data and they are of course accessible by root ... 1115426393 M * Bertl ah, evening terr! 1115426430 M * Doener terr: no problem... as long as we know who's who everything's fine ;) 1115426440 M * terr Hi Herbert. Eager to learn what you uncover about the "init"! 1115426588 M * terr [OT] What IRC client do you use? Which ones would you advise (or warn against)? 1115426623 M * romke Bertl: `quota 500' on host works, on guest not, cat /etc/mtab on guest show no usrquota flag - is this reason why it doesn't work on guest? 1115426644 M * albeiro terr: irssi if you like cmd line. most configurable irc client ever. 1115426708 M * Bertl terr: well, I tested a few .. was relatively unhappy with xchat and used bitchX for some time, then another one (forgot the name) and finally found irssi (and I'm happy with it) 1115426754 M * albeiro yeah :] 1115426763 M * Bertl romke: probably, you should modify the mtab line to contain ufs as filesystem, and have the quota options 1115426786 M * Bertl romke: you also have to use the vroot proxy device if you want it secure 1115426849 M * Bertl terr: add uncover, not much yet, unfortunately I was busy today ... (kind of unexpected) 1115426884 M * albeiro Bertl: or go the linus way ;p 1115426910 M * Bertl terr: but Doener and I are approaching the point where we can actually track down the issue ... 1115426917 M * terr Did other people complain about similar effects? 1115426926 M * romke Bertl: just edit /etc/mtab on guest? 1115426952 M * Bertl terr: currently we have 3 cases which are, hmm .. well somewhat similar .. but not identical 1115426955 M * Bertl romke: yup 1115426991 M * romke Bertl: /dev/hdv1 / ufs rw,usrquota 0 0 1115426995 M * terr All connected to vshelper ? 1115426998 M * romke [/dev/hdv1] 1115426998 M * romke [No quotas on this system] 1115427092 M * Bertl ls -la /dev/hdv1 1115427178 M * romke ls: /dev/hdv1: No such file or directory 1115427233 M * Bertl well, this is supposed to be your vroot device ;) 1115427264 M * romke Bertl: how to make it on guest? 1115427287 M * Bertl you configure it on the host, then simply copy it over ... 1115427308 M * Bertl (best with cp -va) of course you could also create it there 1115427368 M * romke Bertl: mknod i suppose? it should be same address as host device on which this guest is located? 1115427440 M * Bertl no, it should be a vroot device 1115427456 M * Bertl you can use the host device for testing, but that's not secure ... 1115427556 M * Doener ah, compiling inside qemu is fun as always... 1115427602 M * Bertl hmm, not so bad with qemu-fast ... 1115427617 M * Bertl (probably also acceptable with the kernel module) 1115427644 J * eXplasm2 ~explasm@p549FDB1D.dip.t-dialin.net 1115427724 Q * Hollow Remote host closed the connection 1115427777 M * romke Bertl: any Howto on configurig vroot proxy device ? 1115427835 M * Doener $ vrsetup --help 1115427835 M * Doener Usage: vrsetup [-dID] 1115427835 M * Doener or vrsetup 1115427836 M * Doener ;) 1115427848 M * Doener (no idea what -dID is though...) 1115427867 M * Doener maybe the minor number... 1115427930 M * Bertl I guess those are the options ;) 1115427985 M * Doener guess someone has to take a look at the source... 1115428007 M * Bertl vrsetup V0.13 1115428008 M * Bertl options are: 1115428008 M * Bertl -h print this help message 1115428008 M * Bertl -d delete (unset) vroot device 1115428030 M * terr Any hint as to why so few people have encountered the problem with vshelper? 1115428037 M * Doener my vrsetup is from alpha util-vserver... 1115428042 M * Bertl so I make an educated guess and say, enrico did include my debug options to inc and dec the module count too ;) 1115428054 M * Doener ah ok... 1115428064 M * Bertl so -d is probably delete, -I -D is to be ignored ;) 1115428072 Q * eXplasm Ping timeout: 480 seconds 1115428106 M * Bertl terr: probably because it just happens under certain conditions ... 1115428133 M * terr Any idea which ones? 1115428139 M * Bertl terr: using plain init style and most recent vserver versions seems to be some kind of common denominator ... 1115428187 M * terr To have its own "init", plain is the way; right? 1115428258 A * Doener still wonders how init gets into uninteruptible sleep... 1115428510 M * terr How can you track down the problem? 1115428543 M * Doener I'll try my usual wild guessing + cscope ;) 1115429292 M * romke Blah, I'm too tired to get this vrsetup right 1115429296 M * romke nite all 1115429317 M * Doener night romke 1115429413 M * Bertl night! 1115429671 M * albeiro night romke :] 1115429760 M * Bertl hmm, does anybody remember the tool to set the 'main' console in linux? 1115429780 M * Bertl (i.e. to enable kernel messages for a certain console) 1115429784 M * Doener 'main' console? 1115429795 M * Doener ah... 1115429820 M * albeiro i know what are you speaking about but... hm 1115429903 M * albeiro dmesg can only set level 1115429934 J * Doener` ~doener@p54873C26.dip.t-dialin.net 1115430089 Q * Doener Read error: Operation timed out 1115430293 M * Bertl hmm, seems there just _is_ no tool which calls set_console() yet :/ 1115430439 M * terr http://lunar-linux.org/?q=node/view/781 1115430441 M * Doener` http://lunar-linux.org/?q=node/view/792 1115430451 M * Doener` hmm... :) 1115430452 M * terr Is this what you mean? 1115430483 M * terr They talk about setting it in syslogd.conf 1115430485 M * albeiro gentoo booted :] 1115430493 A * Bertl looking ... 1115430510 M * Doener` albeiro: well... that have been 2 hours ;) 1115430541 M * albeiro yeah, but with rest of backup to be done, networking problems and so on 1115430545 M * albeiro not bad thoguht ;] 1115430588 M * albeiro btw, it feels much faster on ext3 :] 1115430619 M * Bertl hmm, I guess you emant the dmesg -n1 no? 1115430638 M * Doener` plus the syslog.conf entry... 1115430654 M * Bertl well, I don#t want to get it from syslog 1115430671 M * Bertl that's the wrong way ... 1115430689 M * terr Why? What's the good one? 1115430697 M * Bertl register_console - take on kernel console duties 1115430700 M * Bertl extern void register_console(void(*proc)(const char*)) 1115430704 M * Bertl This function registers with the kernel a function to dis­ 1115430704 M * Bertl play all the kernel messages printed by the printk(9) 1115430704 M * Bertl function. 1115430735 M * Bertl writing a tool now ;) 1115430764 M * albeiro and publishing it later (it will be simple one) ;] 1115430868 M * Bertl no I will sell it ... NOT 1115430885 M * albeiro i knew you are evil deep inside ;p 1115430890 M * daniel_hozac and get rich! everyone and their dog will want it. 1115430901 M * daniel_hozac ;) 1115430905 M * Bertl yeah, and I'll charge you royalties ;) 1115431031 M * Doener` ok, i got the uninterruptible init... but no zombies... 1115431071 M * Bertl hey great! 1115431167 M * Doener` that's not fair... i want zombies too, mum! 1115431172 M * Doener` ;) 1115431204 M * Bertl hmm ... try to kill the init (with -9 ) 1115431242 M * Bertl and I can understand that, it's not the same -without_ zombies ;) 1115431317 M * Doener` uhm... how should i kill init? it's protected... 1115431324 M * Bertl vkill? 1115431331 M * Doener` it's in context 0... 1115431345 M * Bertl vkill? 1115431356 M * Doener` gets me -ESRCH 1115431358 M * Bertl but probably doesn't help ... 1115431378 M * Bertl ah, we check that? on purpose? 1115431404 M * Doener` vxi = locate_vx_info(id); 1115431404 M * Doener` if (!vxi) 1115431404 M * Doener` return -ESRCH; 1115431428 M * Bertl k, good! 1115431467 M * Doener` but why do i not get zombies? 1115431543 M * Bertl probably ebcause _your_ init is still reaping? 1115431575 M * Doener` hmm... doesn't uninterruptible mean 'i don't do anything till X' 1115431636 M * Doener` ok, old context went away when i entered an other context 1115431656 M * Doener` another start/stop -> zombies in vserver... it's a start ;) 1115431670 M * Bertl damn! my genious masterplan will not work ... 1115431689 M * Bertl (there is no exported interface to change the console after bootup) 1115431853 M * Bertl Doener`: I guess we should 'process' the helper return code like a 'normal' process exit code, no? 1115431970 M * terr http://www.ftlinuxcourse.com/FTLC_Complete/ExternalContribs/FTeBooks/Linux_Device_Drives2/ch04.html 1115431984 M * terr Look for "setconsole" 1115432022 M * terr Maybe it's what you want... 1115432077 M * Bertl ahh, yes, it changes the flags directly ... 1115432543 M * Bertl unfortunately it doesn't help here, as I want to activate the serical console .. but no problem right now ... 1115432607 M * Doener` Bertl: it's a hanging vshelper... 1115432627 M * Bertl yup, hmm, didn't I mention that? 1115432639 M * Doener` not that i remember... 1115432639 M * Bertl the question is, why is it hanging? 1115432654 J * hillct ~hillct@client200-5.dsl.intrex.net 1115432672 M * hillct evening all! 1115432678 M * Bertl evening hillct! 1115432684 M * Bertl Doener`: sorry then ... 1115432694 M * hillct hi Bertl 1115432721 M * hillct what does it take to rename a vserver as in the directories (config tree and root dir) ? 1115432737 M * hillct where do I need ot make the changes in the config tree? 1115432746 M * hillct and is it worth it? 1115432759 M * daniel_hozac it's just a few symlinks. 1115432770 M * daniel_hozac 3 or so in my trees, IIRC. 1115432781 M * hillct I just don't want to screw it up 1115432782 M * Doener` at least i know now how to get rid of the sleeping init without rebooting :) 1115432797 M * terr How ?? 1115432804 M * Doener` kill the vshelper... 1115432818 M * Doener` (just a workaround) 1115432892 M * hillct daniel_hozac can you point me to the proper symlinks? 1115432989 M * hillct hmm. I suppose another approach might be to create a new vserver and just copy over the content in the root of the existing one 1115432996 M * hillct is that a safer method? 1115433010 M * Bertl good idea ... renaming .. we should add that too ... 1115433030 M * hillct ? 1115433041 M * hillct are we listing things that are in need of documentation? 1115433047 M * hillct ;) 1115433222 M * Bertl well, need of documentation _and_ implementation ;) 1115433265 M * hillct ah 1115433403 M * hillct I still intend to write the webmin module for vhost management, and with the file based config tree (which, BTW is really growing on me) it should be ralatively easy to perform a lot of these basic operations 1115433490 M * Bertl sounds good ... 1115433533 M * hillct but i still need to know whay symlinks i need to change in order to rename a vserver ;) 1115433888 M * Bertl I would suggest the following approach: 1115433916 M * Bertl find -type l /vservers/ 1115433925 M * Bertl and then 1115433941 M * Bertl find -type l /vservers/ -exec ls -la {} \; 1115433969 M * Bertl after 'looking' at the output, you should 'know' what to adjust, no? 1115433999 M * hillct yah 1115434021 M * Bertl how do I configure a point to point interface in debian? 1115434035 M * Doener` Bertl: looks like a race between the restart and the shutdown vshelper calls 1115434096 M * Bertl well, I assume (out of the blue) that the restart vshelper 'causes' the shutdown, and doesn't exit afterwards ... 1115434141 M * Doener` there's only one state file, and this contains the name of the pipe the synchronization thing uses 1115434178 M * Bertl btw, what do we need the synchronization thingy for? 1115434184 M * Doener` i don't know ;) 1115434201 M * Doener` i didn't even find out where it is actually started... 1115434204 M * Bertl okay, in any case, elt's file a bug report on savannah ... 1115434470 M * Bertl ah, figured the p2p issue ... 1115434524 Q * anon1 Quit: 1115434891 M * Bertl btw, anybody interested in digging a little through docu and checking what might be usable? 1115435113 M * hillct daniel_hozac is correct. there are 3 links that must be reset 1115435137 M * hillct any preference as to on what page this information should be placed? 1115435207 M * Bertl probably we should put it on a 'TweakingConfig' page, no? 1115435232 M * hillct sure 1115435240 A * hillct browses for the proper page 1115435283 M * Bertl http://linux-vserver.org/TweakingTheConfig (for example) 1115435360 M * hillct ah. I gifured there was a page where such things were already being collected 1115435381 A * hillct starts the page; feels powerful ;) 1115435413 M * Bertl yeah, add a reasonable structure, so that folks intuitively 'know' where to add what ... 1115435424 M * Bertl make sure to set your name and use the preview feature ;) 1115435427 A * hillct tries to remember Wiki syntax 1115435444 M * Bertl (the fine thing is you can _look_ at other pages ;) 1115435494 M * hillct yah 1115435511 M * hillct still, I should have memorized at least some of this crap by now 1115435518 A * hillct is not a great fan of Wikis 1115435532 A * Bertl .o( crap is he saying ... ) 1115435917 M * Bertl Doener`: why does the shutdown vshelper call do anything at all? 1115435937 M * Bertl (just managed to really comprehend what you said) 1115436026 M * Doener` i'm _really_ happy that i could figure out anything... bash hell! ;) 1115436055 M * Doener` it seems to read some information from the state file, which is created by "vserver foo stop" 1115436070 M * Bertl still? 1115436083 M * Bertl didn't I already submit a patch for that twice? 1115436086 M * Doener` and takes that into consideration when deciding which action to take 1115436102 M * Bertl that's plain stupid for startup and shutdown ... 1115436115 A * Bertl is going to look for that patch 1115436147 M * Bertl http://vserver.13thfloor.at/Experimental/NGNET/delta-vshelper-delegate.diff 1115436147 M * Doener` i have vshelper log output of 2 stops... one where the race happens so that state is already deleted -> vshelper shutdown bails out 1115436158 M * Bertl is that in the vshelper now? 1115436173 M * Doener` and one where the race happens so that both, shutdown and restart get the state file -> hanging vshelper 1115436213 M * Doener` yep, that is applied 1115436253 M * Bertl okay, so 'defining' such helpers should solve the issue, no? 1115436280 M * Bertl (which was what I somehow expected to happen .. but obviously didn't 1115436312 M * Doener` checking... 1115436336 M * Doener` but it should, as it would never reach the relevant stuff 1115436396 M * Bertl btw, with the vwait, the restart operation schould spawn a waiting process to do the actual start, once the context is gone ... 1115436513 M * Doener` yep 1115436549 M * Bertl although I'd prefer something like the rebootmgr, but more some kind of 'control' daemon ... 1115436564 M * Bertl (which might even run with lower caps) 1115436593 M * Bertl but that's something for the future ... 1115436613 M * terr Tomorrow? ;-) 1115436664 M * Bertl yeah, maybe ;) 1115436675 M * Doener` yup, works fine if a delegate script is there... (5 test run, previously at least every second one resulted in stuck vshelper) 1115436704 M * Bertl okay, let's do a field test, if anybody is still awake ... 1115436712 M * Bertl terr: you're awake? 1115436731 M * terr Yes. 1115436744 M * Bertl excellent ... listen to Doener`, what to do ... 1115436751 M * terr Was about to leave, but tell me. 1115436814 M * Doener` mkdir -p /etc/vservers/.defaults/apps/vshelper-delegate 1115436820 M * Doener` touch /etc/vservers/.defaults/apps/vshelper-delegate/shutdown 1115436823 M * Doener` chmod +x /etc/vservers/.defaults/apps/vshelper-delegate/shutdown 1115436902 M * hillct http://linux-vserver.org/TweakingTheConfig 1115436913 M * Bertl hmm, maybe echo "#!/bin/true" >/etc/vservers/.defaults/apps/vshelper-delegate/shutdown 1115436957 M * Doener` Bertl: *LOL* actually the restart handler does _not_ use synchronization... resulting in: 1115436961 M * Bertl hillct: looks good, except for the slightly confusing >-> 1115436962 M * Doener` + exec /usr/local/sbin/vserver --defaulttty /usr/local/etc/vservers/test1 restart 1115436962 M * Doener` vserver 'test1' already running 1115436986 M * Doener` a few seconds later the vserver is gone, and the restart thus turned into a poweroff ;) 1115437007 M * Bertl sounds familiar ... *sigh* 1115437050 M * Bertl anyway, we'll fix this up ... 1115437083 M * Doener` yeah, but not today... i'm tired... ;) 1115437105 M * Bertl np, you did a good job today in figuring that stuff out ... 1115437113 M * Doener` good night folks! 1115437118 M * Doener` thanks! :) 1115437122 N * Doener` Doener_zZz 1115437132 M * Bertl night Doener_zZz! 1115437136 M * terr "vserver start/stop" 1115437144 M * terr It works, no? 1115437154 M * terr Here it seems it does! 1115437157 M * Bertl it should work ... 1115437173 M * terr No 'D' state anymore for init 1115437184 M * Bertl good! 1115437206 M * terr Could you explain the magic? 1115437224 M * Bertl it's basically so that the vshelper is called on different occasions 1115437243 M * Bertl e.g. reboot, startup, shutdown, 1115437268 M * terr What is vshelper-delegate ? 1115437272 M * Bertl now the vshelper does some kind of 'synchronization' we currently do not know why ... 1115437309 M * Bertl but, because the vshelper should not look at the startup/shutdown messages for now (it's something for ngnet) 1115437332 M * Bertl this 'additional' synchronization leads to a deadlock 1115437346 M * Bertl (because two calls of a single vserver race) 1115437377 M * Bertl adding the 'dummy' delegate 'avoids' the sync code, so the race can not happen 1115437412 M * Bertl this also explains why I could not recreate those issues ... 1115437425 M * terr Why? 1115437428 M * Bertl the delegates are defined in my setup ;) 1115437436 A * hillct heads off to migrate more websites 1115437439 M * terr Ooh! 1115437444 M * hillct hanks! 1115437447 M * hillct thanks! 1115437449 P * hillct 1115437495 M * terr What are those 'delegates' used for? 1115437520 M * Bertl with ngnet, userspace has to setup the virtual interfaces 1115437531 M * Bertl right before the vserver context starts .. and remove them 1115437545 M * Bertl jsut after the last process exited (startup/shutdown) 1115437598 M * terr And ngnet is not included in the vs2.0 ? 1115437620 M * Bertl no, we decided against it ... too experimental 1115437655 M * Bertl but that alsoexplains why it didn#t happen with older releases, because some of them did not have the helper callbacks 1115437728 M * terr So ngnet is another patch, on top of vs ? 1115437827 M * terr What's its main novelty wrt current ? 1115437929 M * Bertl it introduces completely virtualized networking (next generation networking) 1115437950 M * Bertl you get your personal lo/eth0/whatever in each vserver 1115438033 M * terr The main point is hiding traffic from the host/other vservers ? 1115438070 M * Bertl well, the main point is satisfying some tools (and most users) with proper ifconfig/route/etc output ;) 1115438118 M * terr That's nice of you ;-) 1115438131 M * terr How can I enable it? 1115438146 M * Bertl you need the ngnet patch, then it's enabled 1115438167 M * terr To be applied after vs ? 1115438173 M * Bertl but as I said, experimental, so you need to do some things manually 1115438187 M * Bertl yes, ontop of vs* patches 1115438207 M * Bertl any debian folks around/awake? 1115438218 M * terr Is it explained somewhere? 1115438236 M * Bertl well, there is a howto IIRC 1115438269 M * terr OK, I'll try and find it... 1115438281 M * Bertl http://linux-vserver.org/NGNET-Testing-HOWTO 1115438284 M * Bertl http://linux-vserver.org/ToDo+List+NGNet 1115438293 M * terr Thanks. 1115438309 M * Bertl you're welcome! might be already outdated in some points 1115438361 M * terr So you are using Debian now? 1115438446 M * terr BTW, the vshelper workaround will be included in 0.30.208 ? 1115438452 M * Bertl well, I have to ... on some archs ... 1115438492 M * terr And which is your distribution of choice ? 1115438547 M * Bertl my self created mandrake hybrid ;) 1115438670 M * terr OK... I'm going to sleep now. 1115438702 M * Bertl k, have a good night and thanks for your time! 1115438711 M * terr So I wish you a good night (well, day!) 1115438751 M * terr Thanks to you; I wish I could be of more help! 1115438778 M * Bertl well, reporting issues _is_ important 1115438803 M * Bertl we can not fix what we do not know of ... 1115438826 M * terr Ah, I feel good now, thanks a lot :-) 1115438848 M * terr Bye, bye. 1115438891 P * terr 1115446551 M * Bertl night folks, I'm off to bed now ... 1115446564 N * Bertl Bertl_zZ 1115457745 J * Pazzo ~Pazzo@host130-250.pool8172.interbusiness.it 1115458010 Q * Pazzo Quit: 1115461108 J * Hollow ~Hollow@home.xnull.de 1115465320 M * ntrs Anyone here? 1115465353 M * ntrs something strange has happened 1115465365 M * ntrs cd /vservers 1115465370 M * ntrs touch test 1115465370 M * ntrs touch: cannot touch `test': Read-only file system 1115465411 M * ntrs mount shows that the /vservers filesystem is rw, not ro 1115465495 M * ntrs anyone? 1115468971 M * mountie lsattr 1115468999 M * mountie My guess is that the install made it 000 and chattr'ed it. 1115469060 M * mountie It does that so if any attempts to get out of the chroot in some way succeed, it can't get up into the real root fs. An other layer of security 1115469378 M * FaUl hmm? 1115469381 M * FaUl oh, ww\ 1115469569 J * prae ~prae@sherpadown.net 1115471397 M * Snow-Man Erm, unless you're roto. 1115471398 M * Snow-Man err, root 1115473584 J * monrad ~monrad@213083190130.sonofon.dk 1115473697 Q * prae Quit: Pwet 1115473726 J * hellekin ~hellekin@v41.ath.cx 1115474308 M * mountie Snow-Man: Even if your root, it's much more difficult. you can't follow .. and chroot out again. and you don't have caps to un -i it. 1115474331 M * mountie But ya, nothings perfect. Perfect is the enemy of good 1115474370 M * Snow-Man Be nice if we didn't have to depend on that. 1115477001 Q * monrad Ping timeout: 480 seconds 1115478467 N * Doener_zZz Doener 1115480325 N * Bertl_zZ Bertl 1115480331 M * Doener morning Bertl! 1115480338 M * Bertl morning Doener! 1115480407 M * Bertl mountie: hmm, please elaborate ... 1115480436 M * Bertl ntrs: do not trust the output of 'mount' ever ... 1115480708 M * Bertl Doener: already a patch for rc4? 1115480736 M * Doener hey, i didn't even have a coffee yet! ;) 1115480752 M * Doener but i'm on my way... 1115480759 M * Bertl hmm, ah yes, you just got up too, sorry ;) 1115480791 M * Doener still need to figure out why the syslog stuff crashes on my rc3 patch though... but maybe rc4 makes that go away on its own 1115481001 M * Bertl hmm, wasn't that the 'permission' you added? 1115481012 M * Doener hm? 1115481055 M * Doener when we tested the syslog stuff on my .12-rc3-whatever we got oopses and IIRC even a panic... 1115481073 M * Doener on .11.7-same_whatever it worked flawlessly 1115481118 M * Bertl ah, forget it, I read -pre3 vs -pre4 1115481198 M * Doener hu? what? did you say anything? ;) 1115481288 M * Bertl is linux-vserver.org down? 1115481492 M * Doener looks like... 1115482512 M * Bertl okay, off for dinner now .. back in half an hour ... 1115482530 N * Bertl Bertl_oO 1115483853 M * Doener i'm also off for dinner... back in an hour i guess 1115483857 N * Doener Doener|gone 1115484267 N * Bertl_oO Bertl 1115484792 J * mep_ mep@p5091F2CC.dip.t-dialin.net 1115484975 M * Bertl welcome mep_! 1115485229 Q * mep__ Ping timeout: 480 seconds 1115485816 N * mep_ mep 1115485818 M * mep thx :) 1115486150 J * Nik ~Nik@cable-153-130.online.bg 1115486153 M * Nik hi all 1115486270 M * Bertl evening Nik! 1115486277 M * Nik Bertl :-) 1115486323 M * Nik Bertl: i'm playing with the 2.0-pre4 and have a problem... 1115486334 M * Nik Bertl: this one stucks: /bin/bash /sbin/vshelper shutdown 10001 1115486346 M * Nik Bertl: while trying to stop a vserver 1115486363 M * Bertl yup, bug in the tools, you are subscribed to the ml? 1115486385 M * Nik Bertl: no, but I guess i schuld have been ;-) 1115486403 M * Bertl sec, looking for a copy ... 1115486409 M * Nik Bertl: allready patched? 1115486475 M * Bertl http://www.paul.sladen.org/vserver/archives/200505/0085.html 1115486523 M * Nik Bertl: 10x :-) 1115486586 M * Bertl you're welcome! 1115487338 M * romke Bertl: it's for my issue? great :> 1115487569 M * Bertl yup! 1115487641 Q * Nik Ping timeout: 480 seconds 1115487642 M * Bertl should solve the issues you had as well as pazzos recent discoveries 1115488248 M * romke :> 1115488254 M * romke will try 1115488285 J * Jani ~Jani@G9db0.g.pppool.de 1115488287 M * Jani Hi all 1115488318 M * Bertl welcome Jani! 1115488326 M * Jani *hmms* Homepage is down? 1115488335 M * Bertl yup, seems so ... :( 1115488357 A * Jani is innocent ;) 1115488397 M * Bertl any proof for that? *G* 1115488471 M * Jani *hides the proofs* 1115488488 M * Jani So, 3. vServer is goind online today :) 1115488528 M * Bertl excellent! which kernel/patches? 1115488581 M * Jani Kernel 2.6.11.5-vs1.9.5 1115488597 M * Bertl everything fine? 1115488696 M * Jani ! 1115488703 M * Jani Sofar yes :) 1115488711 M * Bertl good! 1115488714 M * Jani Sometimes it's a little to easy ;) 1115488725 M * Bertl ah, we have patches for that ;) 1115488883 M * Jani *g* 1115489070 M * Jani Btw. have someone experience with WebDAV? 1115489111 M * Bertl not really ... 1115489962 J * monrad ~monrad@213083190130.sonofon.dk 1115490130 N * Doener|gone Doener 1115490140 M * Doener evening folks 1115490192 M * Jani *waves* 1115490611 M * Bertl wb Doener! 1115492227 M * Bertl today any debian folks present? I want to know how to 'run' a script when an interface is to be taken up ... 1115492364 M * albeiro there is ifcfg used 1115492392 M * albeiro with scripts executed in each state. do not ask me about details ;] 1115492417 M * Bertl hmm, yeah, figured that, but I need to run a script before if* can work ... 1115492467 M * Bertl but there seems to be some 'script' hook in interfaces ... investigating now 1115492487 M * Bertl yup pre-up command 1115492488 M * albeiro yeah. and there should be something ending with .pre 1115494361 J * Nik ~Nik@cable-153-130.online.bg 1115494846 Q * Nik Ping timeout: 480 seconds 1115494973 J * Doener` ~doener@p5487480E.dip.t-dialin.net 1115495020 M * Doener` stupid router... 1115495126 Q * Doener Read error: Operation timed out 1115497979 J * steve ~steve@user-2511.lns4-c8.dsl.pol.co.uk 1115497997 M * steve hey all 1115497998 N * steve Guest39 1115498004 N * Guest39 steve- 1115498020 M * Bertl hey steve*! 1115498021 M * steve- how do I build a vserver which just copies the files from the root server? 1115498026 N * steve- Guest40 1115498038 M * Guest40 I've looked at the docs but can't really find anything 1115498044 N * Guest40 steve--` 1115498050 M * Bertl vserver --help 1115498067 M * steve--` yeah I did that 1115498073 M * steve--` but "build *" - what are the buildops? 1115498095 M * Bertl vserver - build --help 1115498110 M * steve--` aha! 1115498247 M * steve--` hmm 1115498248 M * steve--` /usr/local/lib/util-vserver/vserver-build: line 125: /usr/local/lib/util-vserver/vserver-build.copy: No such file or directory 1115498280 M * Bertl which tool version? 1115498319 M * steve--` util-vserver-0.30.204 1115498321 M * steve--` for the 2.6 kernel 1115498328 M * Bertl try to update to 0.30.207 1115498377 M * Bertl but I'm not sure it is there ... 1115498397 M * Bertl (if it is not, then you should file a bug report) 1115498501 M * steve--` when I try and install from the RPM package it gives this failed dep 1115498502 M * steve--` error: Failed dependencies: 1115498502 M * steve--` init(util-vserver) is needed by util-vserver-0.30.204-1mdk.i586 1115498508 M * steve--` any idea where I can get init(util-vserver) from? 1115498550 M * Bertl are you using mandrake? 1115498564 M * steve--` centos 1115498578 M * steve--` ah, mdk means mandrake? 1115498613 M * Bertl yep, that's the mandrake extension 1115498616 M * steve--` ah 1115498626 M * steve--` do you know where I can get a centos RPM from? I couldn't find any 1115498627 M * Bertl it's probably best to build it from the tar with 1115498630 M * steve--` ah 1115498641 M * Bertl rpm(build) -tb 1115501903 M * gabaZzzzleeping hi 1115501907 N * gabaZzzzleeping gaba 1115501978 M * gaba i have a question: i have kernel 2.6.10 and i see that the last download for the patch in http://www.13thfloor.at/vserver/s_release/v1.2.10/ says Kernel Release 2.4.30. That means that the patch is for the kernel 2.4.30 ? 1115502080 M * daniel_hozac http://66.102.9.104/search?q=cache:f_pUZrwOS5oJ:linux-vserver.org/Release%2BFAQ+&hl=en&start=1 1115502144 M * gaba mmm, and i dont find a patch for 2.6.10 1115502158 M * daniel_hozac 2.6.10 is too old ;) 1115502177 M * daniel_hozac http://vserver.13thfloor.at/Experimental/ 1115502218 M * gaba old? 1115502241 M * gaba ahh, this is experimental.. :) 1115502274 M * gaba here: http://www.13thfloor.at/vserver/e_patches/vs-26x/ all of them are 2.6.0... 1 to 3 1115502429 M * Bertl gaba: why do you want to use 2.6.10? 1115502556 M * gaba is what i have in the server 1115502640 M * daniel_hozac and why wouldn't you want to upgrade to the latest and greatest? 1115502650 M * gaba mmm 1115503137 M * gaba maybe but i would prefer not to do it 1115503140 M * gaba just install the patch 1115503157 M * Bertl aha, because of the 'known' bugs? 1115503164 M * gaba known bugs? 1115503171 M * Bertl well, in 2.6.10 ... 1115503360 M * Bertl nevermind, was just curious ... 1115507214 M * Doener` Bertl: any plan for tonight? 1115507643 M * Bertl hmm, I'm already working on a few things .. but if you are interested in something special .. I'll gladly join this too ;) 1115507665 M * Bertl (currently running things do not need much of my attention anyway) 1115507826 M * Doener` i'm searching for sth. to do ;) i've postponed hacking on util-vserver for a day or two... had too much bash yesterday ;) 1115507850 M * Bertl what about hacking on ngnet then? 1115507918 M * Bertl or maybe adding a flag to disable the startup/shutdown calls? 1115507920 M * Doener` hmm... i've spent a lot of time thinking about that arp stuff... even about implementing it some other way (the whole thing, not arp)... but no solution showed up in my brain 1115507945 M * Bertl ad arp: IMHO the following should work quite fine: 1115507956 A * Doener` .oO( maybe i should stick an invitation for ideas onto my forehead... ) 1115507970 M * Bertl - make a global 'waiting for arp reply' list 1115507996 M * Bertl - add the arp requests to that (as already done in linux) 1115508027 M * Bertl - when replies arrive, filtem them and present them to the 1115508036 M * Bertl 'waiting' contexts ... 1115508065 M * Bertl - for outgoing arp requests, again filter them, and if permitted 1115508079 M * Bertl generate fake arp replies for local addresses 1115508121 M * Bertl but, I have another option for that too ... 1115508176 M * Bertl we could move this 'policing' out of the kernel and do some kind of 'arp-daemon' which handles requests _and_ replies from userspace 1115508219 M * Bertl maybe via netlink? 1115508329 M * Doener` hmm... i'll take another deeeeep look at the code... 1115508330 M * albeiro second one could be more extensible 1115508344 M * albeiro sorry for interupting btw ;] 1115508361 M * Doener` input is always welcome :) 1115508370 M * Bertl np, yeah, actually I'd prefer several such helper daemons on the host ... 1115508400 M * Bertl we have a lot of policy in the kernel, where it really doesn't belong IMHO 1115508414 M * albeiro indeed, policy is userspace thing 1115508427 M * albeiro and framework belongs to the kernel 1115508428 M * Bertl e.g. I could imagine the entire proc security thing to go to userspace ... 1115508469 M * daniel_hozac won't the outsourcing to userspace slow it down though? 1115508512 M * Bertl very like ... but a) it could be cached, and b) do we really need ulta-fast proc? 1115508536 M * albeiro caching is tricky ;] 1115508565 M * Bertl well, for the proc, it really isn't i.e. there is the dentry and the inode 1115508589 M * Bertl as long as the dentries are not flushed, the relevant info is cached by default 1115508604 A * albeiro recalls dazuko caching in rsbac. ok, it was all coded in few hours. caching is not tricky ;p 1115508616 M * Doener` hm... caching already causes a small problem there (sometimes wrong information for init in context 1) 1115508675 M * Bertl Doener`: do you have a test case for that, I still consider it a minor bug on our side ... 1115508697 M * Doener` sec.. 1115508835 M * Doener` http://www.13thfloor.at/~doener/init-bug 1115508869 M * Doener` as you see, the host's init is "replaced" by the context fake init process 1115508884 M * Bertl yup, sec checking/looking at the code ;) 1115509253 M * steve--` when you say 'context' what exactly do you mean? 1115509299 M * Bertl the linux-vserver concept is based on 'contexts' which are isolated against each other (process/resource wise) 1115509324 M * steve--` ah 1115509339 M * Bertl if you look at the url Doener` pasted, you probably get an idea 1115509343 M * steve--` yeah 1115509362 M * steve--` I got the vserver created in the end, though when I try ps it says it's not mounted 1115509363 M * steve--` any ideas? 1115509368 M * steve--` [root@server1 /]# ps ux 1115509369 M * steve--` Error: /proc must be mounted 1115509371 M * Doener` http://linux-vserver.org/Proc-Security 1115509375 M * steve--` ah 1115509380 M * Bertl oops .. have to cleanup ... sec 1115509401 M * steve--` site seems to be down for me 1115509407 M * Doener` ah right... 1115509422 M * Bertl yup, unfortunately still is .. hopefully jack will respond soon ... 1115509426 M * Doener` http://64.233.161.104/search?q=cache:HX4YokE9kToJ:linux-vserver.org/Proc-Security+proc+security&hl=de 1115509444 M * Bertl yeah, google should have a view copies ... 1115509465 M * Doener` that should be more readable (no highlighting) http://64.233.161.104/search?q=cache:HX4YokE9kToJ:linux-vserver.org/Proc-Security 1115509838 M * Bertl okay, now rebuilding ... vs2.0-pre4 1115510306 M * Bertl fs/proc/base.c 1115510329 M * Bertl line 1115, right before the 'discard wrong fakeinit' comment 1115510344 M * Bertl remove the VX_WATCH from the vx_check() 1115510392 M * Bertl basically we should add a check for the rmapped pid there ... but that is maybe something you want to do? ;)