1155772803 M * Bertl derjohn: ping!
1155772822 M * daniel_hozac none of the privacy/disol/etc. patches?
1155772828 M * Bertl note, unless somebody objects till the weekend, I consider the 2.0.2 final
1155772857 M * daniel_hozac sounds like a plan.
1155772881 M * cehteh mhm
1155772882 M * Bertl - yeah, privacy and friends are there in devel
1155772912 M * cehteh i am just reinstalling my new server .. how about making it final now? :)
1155772937 M * Bertl I doubt that somebody will object :)
1155772965 M * Bertl (so you can make it final for you _now_ and it will probably become final on the weekend :)
1155772973 M * cehteh :P
1155773040 M * cehteh anyways .. i am just finished with the hardware, cleaning, new fans ... have to instal/partiton it before building a vserver kernel .. maybe tomorrow
1155773061 M * Bertl okay, off to bed now ... have fun everyone! tx, and cya!
1155773069 N * Bertl Bertl_zZ
1155773070 M * cehteh cu ;)
1155773108 Q * coocoon Quit: KVIrc 3.2.0 'Realia'
1155775164 J * harry ~harry@d54C2508C.access.telenet.be
1155776305 Q * pflanze Ping timeout: 480 seconds
1155776860 J * shedii ~siggi@inferno.lhi.is
1155777073 Q * harry Ping timeout: 480 seconds
1155777224 Q * shedi Ping timeout: 480 seconds
1155777625 M * cehteh       Currently, Linux supports LINEAR md devices, RAID0 (striping), RAID1 (mirroring), RAID4 and RAID5.
1155777625 M * cehteh        Recent kernels (2002) also support a mode known as MULTIPATH.  mdadm does not support MULTIPATH as yet.
1155777662 M * cehteh ... some indication that you have to update a server is when you have to read the manpages because the commands there are to old :)
1155777781 J * Nam ~nam@S0106001195551ff0.va.shawcable.net
1155778094 J * harry ~harry@d54C2508C.access.telenet.be
1155780511 Q * mkhl 
1155781384 Q * Hurga Remote host closed the connection
1155781916 J * FireEgl FireEgl@2001:5c0:84dc:1:4::
1155782268 Q * Karmek Ping timeout: 480 seconds
1155783899 J * ntrs__ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com
1155783900 Q * ntrs_ Read error: Connection reset by peer
1155784069 J * ntrs_ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com
1155784069 Q * ntrs__ Read error: Connection reset by peer
1155786840 J * matti matti@linux.gentoo.pl
1155788480 J * prae ~benjamin@foxhound.sherpadown.net
1155788743 Q * prae 
1155789386 J * Viper0482 ~Viper0482@p54977E41.dip.t-dialin.net
1155790995 J * yarihm ~yarihm@whitehead2.nine.ch
1155791072 J * zkbrsnie ~zkbrsnie@83-64-146-226.klosterneuburg.xdsl-line.inode.at
1155792060 Q * Viper0482 Remote host closed the connection
1155792307 Q * yarihm Quit: This computer has gone to sleep
1155793321 Q * olilo Quit: brb
1155793363 J * olilo hiddenserv@tor.noreply.org
1155793399 J * yarihm ~yarihm@whitehead2.nine.ch
1155794818 Q * Hollow Quit: Konversation terminated!
1155795016 J * Hollow ~hollow@2001:a60:f026::1
1155795354 J * transacid ~transacid@e182078140.adsl.alicedsl.de
1155795474 Q * transaci1 Ping timeout: 480 seconds
1155795715 M * sid3windr mmkay
1155795728 M * sid3windr anyone running clamav in default vserver config may have a problem ;)
1155795738 M * sid3windr as the new datafile doesnt fit on a 16M /tmp
1155795774 M * Hollow sid3windr: yep, i noticed this yesterday :p
1155796039 J * Zaki_ ~Zaki@212.118.109.136
1155796430 Q * Zaki Ping timeout: 480 seconds
1155797285 J * balbir ~balbir@59.145.136.1
1155798342 J * dna ~naucki@p54BCEA72.dip.t-dialin.net
1155798489 Q * Aiken Ping timeout: 480 seconds
1155799849 J * ybanafa ~ybanafa@82.114.176.203
1155800423 N * Belu_zZz Belu
1155800799 P * ybanafa Leaving
1155802843 Q * ntrs_ Read error: Connection reset by peer
1155803037 Q * brc_ Ping timeout: 480 seconds
1155803160 J * ntrs_ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com
1155804568 J * Aiken ~james@tooax8-089.dialup.optusnet.com.au
1155805908 J * Ben81 ~Ben81@tipi0e.lri.fr
1155807884 J * coocoon ~coocoon@p54A0541D.dip.t-dialin.net
1155808852 Q * shedii Quit: Leaving
1155809019 M * derjohn Bertl_zZ, pong ;(
1155809137 Q * schimmi Ping timeout: 480 seconds
1155809427 J * pisco ~pampel@p508781E9.dip0.t-ipconnect.de
1155811017 Q * coocoon Quit: KVIrc 3.2.0 'Realia'
1155811037 J * lilalinux ~plasma@dslb-084-058-195-153.pools.arcor-ip.net
1155811534 J * coocoon ~coocoon@p54A07889.dip.t-dialin.net
1155812094 Q * Aiken Ping timeout: 480 seconds
1155812767 M * mnemoc sid3windr: yep, i had to change 6 guests to 32m at /tmp yesterday :p
1155812800 M * sid3windr :)
1155812805 M * sid3windr 3 here ;)
1155812868 M * matti ;]
1155812947 M * daniel_hozac derjohn: you left out privacy in the devel changelog, the biggest change for a long time ;)
1155812963 M * derjohn daniel_hozac, lart me !
1155812969 M * derjohn daniel_hozac, and. what is "fh" ??
1155813010 J * lolilol hiddenserv@tor.noreply.org
1155813041 M * daniel_hozac file handle, i thought, but cow-fix07 fixes access(2) and other things using permission() for COW files (used to say -EACCES because the files are immutable).
1155813091 Q * olilo Ping timeout: 480 seconds
1155813158 J * gerrit ~kvirc@dslb-084-060-215-168.pools.arcor-ip.net
1155813571 M * derjohn daniel_hozac, but the fh was written in the brackets so I thought it may not be an filehandle ( maybe some fh-version i dont know abozt)
1155813615 M * derjohn daniel_hozac, what do the privacy friend do?
1155813647 M * derjohn daniel_hozac, wasnt that about private ttys or so? (I didnt follow the discussion)
1155813677 M * daniel_hozac disol-feat01 adds tagging and isolation to dm and loop devices.
1155813747 M * daniel_hozac privacy-feat01 adds a configuration option which makes your guests private, i.e. you can't see them from xid 1 (only the processes are visible).
1155813750 M * derjohn disol-feat01 - is that a friend of the privacy? is there even more?
1155813757 M * derjohn ah, race ;)
1155813788 M * daniel_hozac yep, disol is a friend of privacy's.
1155813911 M * derjohn hm, i do not see the advantage of having the guest "hidden" from xid 1? does that mean, if i unhide certain stuff for guest XY , it cant see stuff from guest XY+1 in /proc?
1155813947 M * daniel_hozac what?
1155813960 M * daniel_hozac guests have never been able to see eachother's stuff in /proc.
1155813975 M * daniel_hozac if they have, that's a serious security bug.
1155813978 M * derjohn well, what is "more private" then before?
1155814007 M * derjohn i mean, who "uses" the spectator context?
1155814019 Q * coocoon Read error: Connection reset by peer
1155814024 M * daniel_hozac the admin.
1155814046 M * derjohn hm, doesnt the admin have xid 0 anyway?
1155814053 M * daniel_hozac he does.
1155814077 M * daniel_hozac but along with the removable VXF_STATE_ADMIN, you can create guests pretty much shielded from the host admin.
1155814083 M * derjohn in which case of use will something be hidden from the admin ?
1155814088 M * daniel_hozac (assuming an untainted kernel :))
1155814126 M * derjohn erm, the idea is tio shield the guest from the admin (root), right?
1155814133 M * daniel_hozac right.
1155814141 M * derjohn but we could do anything with xid 0 rights?
1155814143 J * schimmi ~sts@host82.natpool.mwn.de
1155814155 M * daniel_hozac no, that's the point of the VXF_STATE_ADMIN.
1155814164 M * daniel_hozac once it's removed, you can't really touch the context.
1155814206 M * derjohn hm,  except processes?
1155814214 M * daniel_hozac not even processes.
1155814228 M * daniel_hozac you can see the processes, like with vps.
1155814244 M * daniel_hozac (this is all IIUC, i haven't used privacy yet)
1155814260 Q * pagano Ping timeout: 480 seconds
1155814271 M * derjohn sure, nearly nobody did ;)
1155814289 M * derjohn so, admin cant e.g. enter the guests namespace ?
1155814306 M * daniel_hozac that's the idea.
1155814365 M * derjohn so, one possible use in the future is an admin, that is not root ?
1155814382 M * daniel_hozac hmm?
1155814396 M * derjohn i mean, root can enter the namespace anyway if he likes via xid 0, nor?
1155814410 M * daniel_hozac no, that's what i'm saying.
1155814430 M * daniel_hozac he shouldn't be able to touch the context in any way if it doesn't have VXF_STATE_ADMIN.
1155814449 M * derjohn how can he enter VXF_STATE_ADMIN?
1155814459 M * daniel_hozac ?
1155814471 M * derjohn ( i mean: most of that stuff will be handeled by tools I assume)
1155814546 M * daniel_hozac what stuff?
1155814553 M * derjohn let me rephrase my question: If I enable that feature(s) and start a guest. what can I do with that guest? Of course I can start and stop, but I have no way to enter it anymore? vserver XY-private enter doesnt work then?
1155814576 M * daniel_hozac no.
1155814584 M * daniel_hozac and stop is special cased as well.
1155814605 M * daniel_hozac you can only signal init, if the guest has one of those.
1155814631 M * daniel_hozac only when there are no other processes left will you be able to kill init.
1155814851 P * schimmi Verlassend
1155814947 M * Hollow daniel_hozac, derjohn: i tried it yesterday, you can only send SIGINT to init, you cannot migrate, cannot set flags etc..
1155814964 M * Hollow btw, i got totally rid of persistent with the vshelper approach
1155814972 M * Hollow works really nice
1155814976 M * daniel_hozac cool.
1155814993 M * daniel_hozac even with the no-migrate-on-create?
1155815006 M * Hollow no, i even left out this patch...
1155815023 M * Hollow well, doesn't matter anyway
1155815030 Q * mire Ping timeout: 480 seconds
1155815034 M * derjohn daniel_hozac, Hollow : But all this has to be activated explicitly? I would have a strange feeling in the stomach, if admin counld not kill a guest will all it processes. There a sometimes "stuck ones" in crashed guetss.
1155815046 M * daniel_hozac ah right, vshelper is called during _create.
1155815049 M * Hollow i saw that all worked without persistent as well, so i simply dropped it from creation phase :P
1155815062 M * Hollow but we stil do external setup
1155815094 M * daniel_hozac derjohn: CONFIG_VSERVER_PRIVACY is enabled by default.
1155815115 M * Hollow derjohn: as daniel said, kill is a special case... you can kill contexts
1155815118 M * derjohn it's a complile time option?
1155815121 M * daniel_hozac but VXF_STATE_ADMIN has to be removed explicitly.
1155815128 M * Hollow indeed
1155815136 M * Hollow i think i make a new config option for that.. ;)
1155815147 M * daniel_hozac hehe.
1155815157 M * daniel_hozac derjohn: yes.
1155815179 M * Hollow privacy = compile time, state_Admin = run tim
1155815185 M * Hollow but once removed, cannot be added back
1155815242 M * derjohn Hollow, if I enable that compiletime stuff and use the std utils .210 - will I see any differnt behavior?
1155815278 M * Hollow no
1155815315 M * Hollow privacy has to do with xid = 1 and what you _see_, state_admin is what you can _do_
1155815376 M * Hollow vshelper became a tale teller ;)
1155815408 M * Hollow http://paste.linux-vserver.org/250 ;)
1155815423 J * mire ~mire@110-167-222-85.COOL.ADSL.VLine.verat.net
1155815821 J * coocoon ~coocoon@p54A07889.dip.t-dialin.net
1155815853 M * derjohn daniel_hozac, Hollow: http://linux-vserver.org/ChangeLogDevelExperimental .. ok so?
1155815904 Q * balbir Quit: Leaving
1155815942 M * Hollow yep, sounds ok
1155816029 M * daniel_hozac Hollow: well, it will be different. you won't be able to see everything from xid 1 like you used to.
1155816227 N * lolilol olilo
1155816451 M * daniel_hozac Hollow: rofl, great log messages!
1155816544 Q * yarihm Remote host closed the connection
1155816593 J * yarihm ~yarihm@whitehead2.nine.ch
1155817099 M * derjohn Hollow, I would have expected something like context_4_zZ and context_4_oO ;)
1155817135 M * derjohn hm, maybe that only counts for context 0 ;)
1155817333 M * daniel_hozac Bertl_zZ: ah, that 2.6.17.8-vs2.1.1-rc28.ipv6 compile failure i showed you is due to the .ipv6 part.
1155817549 M * Hollow derjohn: hehehe, would be funny too ;)
1155817897 A * sid3windr wonders what time it is in BMT
1155818308 M * daniel_hozac sleep time :)
1155818410 Q * gerrit Quit: Changing server...
1155821187 M * mnemoc hi, what's the address for vserver's paste service ?
1155821214 M * coocoon mnemoc: http://paste.linux-vserver.org/
1155821217 M * mnemoc thanks
1155821235 M * mnemoc uhm, no file uploading :(
1155821430 M * mnemoc Hollow: http://pastebin.ca/136143 <--- small cleanup, and fix because local can't be used outside functions on bash
1155821461 M * Hollow mnemoc: already fixed ;)
1155821466 M * mnemoc =)
1155821562 M * mnemoc Hollow: not quoting $1 can produce funny results there ,-)
1155821591 M * Hollow mhm..
1155821617 M * Hollow next commit will fix it ;)
1155822139 Q * yarihm Quit: Leaving
1155822638 M * mnemoc Hollow: does `vcc create` work?
1155822669 M * Hollow yep, but you must gunzip/bunzip the tarballs before
1155822685 M * Hollow i will implement some template management methods during the next days
1155822720 M * mnemoc Hollow: where are the templates taken from?
1155822746 M * Hollow $localstatedir/vcd/templates
1155822866 M * mnemoc so $localstatedir/vcd/templates/$template/*.tar ?
1155822872 Q * coocoon Quit: KVIrc 3.2.0 'Realia'
1155822908 M * Hollow $localstatedir/vcd/templates/$template.tar
1155822950 M * Hollow in the future there will be scripts similar to vserver ... build, but the daemon will only use templates
1155822988 J * coocoon ~coocoon@p54A07889.dip.t-dialin.net
1155823028 M * mnemoc what do you think about a $template.sh to perform post-extract actions?
1155823068 M * Hollow vx.create supports template config... if you create a file $localstatedir/vcd/templates/$template.conf it will use the default values found there...
1155823081 M * Hollow it currently supports init style, bcaps, ccaps, and flags
1155823135 M * mnemoc my idea was more related to auto pre-customization of the files inside the newly created guest
1155823166 M * mnemoc but ok
1155823169 M * Hollow well, we could also add post-script support
1155823305 M * mnemoc nice =)
1155823363 M * Hollow the start method could also get script support
1155823370 M * Hollow will have to investigate whats best here
1155823432 J * shedi ~siggi@dsl-og-108-50.du.vortex.is
1155823446 M * mnemoc is it too complex to add .tar.bz2 support ?
1155823457 M * mnemoc having expanded templates is not that nice
1155823488 M * Hollow it should be possible..
1155823528 M * Hollow i was previously using libarchive instead of libtar
1155823546 M * Hollow it supported gzip and bzip..
1155823559 M * Hollow maybe switching back would be an option too
1155823668 M * coocoon Hollow: please do that, or maybe all should be made possible
1155823692 M * mnemoc Hollow: you could use exec() instead of using external libraries :p
1155823714 M * Hollow i'd prefer not doing that
1155823715 M * Hollow ;)
1155823753 M * mnemoc using script 'handlers' it may be easier to support package managers
1155823792 J * Wonka produziert@chaos.in-kiel.de
1155823833 M * Hollow the question is: should the scripts be changable using xmlrpc
1155823989 M * Wonka re. vserver running again, after third hangup without visible cause or log entries this week...
1155823999 M * Wonka built new kernel now
1155824026 M * mnemoc they could just receive by enviroment some usefull information, and you only worry about $template.conf can calling the handler
1155824098 M * Hollow mhm
1155824107 M * mnemoc ugly?
1155824131 M * Hollow no, sounds like a plan
1155824147 M * mnemoc =)
1155824151 M * Hollow ;)
1155824219 Q * shedi oxygen.oftc.net strange.oftc.net
1155824219 Q * pisco oxygen.oftc.net strange.oftc.net
1155824219 Q * harry oxygen.oftc.net strange.oftc.net
1155824219 Q * Nam oxygen.oftc.net strange.oftc.net
1155824219 Q * Johnnie oxygen.oftc.net strange.oftc.net
1155824352 J * shedi ~siggi@dsl-og-108-50.du.vortex.is
1155824352 J * pisco ~pampel@p508781E9.dip0.t-ipconnect.de
1155824352 J * harry ~harry@d54C2508C.access.telenet.be
1155824352 J * Nam ~nam@S0106001195551ff0.va.shawcable.net
1155824352 J * Johnnie ~john@dynamic-acs-24-154-53-237.zoominternet.net
1155824405 J * JimmyGulp ~james@ns0.esagroup.co.uk
1155824531 Q * FloodServ charon.oftc.net helium.oftc.net
1155824692 J * FloodServ services@services.oftc.net
1155824745 M * JimmyGulp Howdy, I have a vserver 1.2.10 setup, and I'm trying to deploy a new vserver on the machine using the "deploy-vserver.sh" script from marlow.dk, however, I can't reach marlow.dk, has it disappeared, and if so, where can I get a copy of the images that used to be there (I'm looking for the fc2 and sarge images)
1155824795 M * daniel_hozac you're aware FC2 is completely EOL, right?
1155824840 M * JimmyGulp er, yeah
1155824870 M * daniel_hozac i can reach marlow.dk just fine.
1155824879 M * JimmyGulp hmm...
1155824903 M * JimmyGulp can't get to it from my easynet line (UK), nor from the colo house in london
1155824928 M * JimmyGulp actually, I can now, how odd
1155824951 M * JimmyGulp never mind :)
1155825496 M * FaUl has anyone tried linux-vserver on a notebook? is there something which avoids the suspend-stuff to work?
1155825901 M * daniel_hozac i'm running my kernels on my laptop, but i never suspend :)
1155825947 M * daniel_hozac there shouldn't be though.
1155826060 Q * Ben81 Quit: Leaving
1155826202 Q * cehteh Ping timeout: 480 seconds
1155826674 J * hallyn ~xa@adsl-75-21-68-95.dsl.chcgil.sbcglobal.net
1155827548 J * cskarby ~cs@195.1.31.69
1155827554 P * cskarby 
1155827563 J * cehteh ~ct@cehteh.homeunix.org
1155828223 M * h01ger FaUl, suspend works here
1155828518 J * michal_ ~michal@www.rsbac.org
1155828522 M * michal_ hey
1155828534 M * michal_ i'm trying to start suse (10.1) based vserver guest
1155828556 M * michal_ issued vserver-build -m skeleton, hand configured all the needed files, copied system and...
1155828562 M * michal_ vshelper.init: can not determine xid of vserver 'suse'; returned value was ''
1155828581 M * michal_ i have some services beeing started in this guest
1155828598 M * michal_ sshd for example - with addr it is bondingg to changed of courtse
1155828716 M * michal_ so i have no idea what to do :]
1155829772 J * stefani ~stefani@tsipoor.banerian.org
1155829965 J * gerrit ~kvirc@dslb-084-060-215-168.pools.arcor-ip.net
1155829990 J * Viper0482 ~Viper0482@p54977551.dip.t-dialin.net
1155830218 M * daniel_hozac michal_: are you sure? do you see the services starting?
1155830395 J * lolilol hiddenserv@tor.noreply.org
1155830444 M * michal_ daniel_hozac: you nkow...i know they are added to the proper runlevel - i cannot see any of them starting
1155830461 Q * olilo Ping timeout: 480 seconds
1155830463 M * daniel_hozac initstyle plain or sysv?
1155830477 M * michal_ err msg from vshelper.init is the first message i can see
1155830494 M * michal_ which one is plani? it's a suse guest
1155830596 M * daniel_hozac yes, but what do you have in /etc/vservers/.../apps/init/style?
1155830609 A * sid3windr pokes Bertl_zZ =)
1155830621 M * michal_ nothing :)
1155830634 M * michal_ my apps/init of that vserver is empty actualy
1155830644 M * daniel_hozac ok, so sysv.
1155830656 M * michal_ probably...
1155830660 M * daniel_hozac so they're all set to start in runlevel 3?
1155830670 M * michal_ rnulevel 5
1155830674 M * michal_ *runlevel 5
1155830681 M * daniel_hozac have you configured that?
1155830688 M * michal_ there is something in 3 also...but i'm startin in 5
1155830694 M * daniel_hozac no you're not.
1155830745 M * michal_ platinum:/vservers/suse# ls etc/init.d/rc3.d/ | wc -l
1155830745 M * michal_ 61
1155830758 M * daniel_hozac ls etc/init.d/rc3.d/S* | wc -l 
1155830761 M * michal_ even if it is trying to start in rl 
1155830763 M * michal_ 3
1155830777 M * michal_ a lot of services to start...ok...even fail.
1155830795 M * derjohn would you mind to anset me Q to a c-lamer: in fs/exec.c there is "sprintf(pid, "%ld", current->tgid);". I think that is within the kernel crashdump addon (default in ubuntu edgy kernel). When compileing i get:  warning: format '%ld' expects type 'long int', but argument 3 has type 'pid_t'. i could obvioulsy simply comment out the sprintf, but what is the right way to make it work, i.e. output with format?
1155830835 M * daniel_hozac it's just a warning.
1155830840 M * michal_ this warning is harmless
1155830848 M * michal_ see at pid_t define
1155830855 M * michal_ s/see/look/
1155830885 M * daniel_hozac well, actually, pid_t is int here.
1155830895 J * bonbons ~bonbons@83.222.36.236
1155830897 M * daniel_hozac so %d would actually be correct.
1155830930 M * daniel_hozac bonbons: hey, you know ipv6 causes compilation failures on 2.6.17.8-vs2.1.1-rc28+?
1155830946 M * michal_ ok...and i still have no idea how to make it working...
1155830985 J * transaci1 ~transacid@e176000215.adsl.alicedsl.de
1155831008 M * coocoon michal_: what happens if u set plain in /etc/vservers/.../apps/init/style
1155831041 M * daniel_hozac michal_: so you have scripts that are set to start? does suse have something like Gentoo where it shits all over itself if some run files still exist and won't start any services?
1155831069 M * michal_ i have no idea :]
1155831085 M * bonbons daniel_hozac: hey, I know, saw it two days ago, but had no time yet to look deeper at it
1155831098 Q * transacid Killed (NickServ (GHOST command used by transaci1))
1155831103 N * transaci1 transacid
1155831154 M * michal_ heh
1155831159 M * michal_ plain initstyle made me...
1155831172 M * michal_ not throw any error
1155831179 M * michal_ "started" vserver
1155831185 M * michal_ at least it thnks so
1155831189 M * michal_ did not start any of app
1155831207 M * michal_ vserver suse enter returns me info about suexec beeing for vserver only
1155831216 M * michal_ i cannot stop that "running" vserver now
1155831217 M * michal_ lol
1155831276 M * derjohn daniel_hozac, I was on the wrong train. Here is the error: fs/exec.c:1584: error: too few arguments to function 'vfs_unlink'
1155831282 M * derjohn caused by.
1155831284 M * derjohn vfs_unlink(parent->d_inode, dentry);
1155831324 M * coocoon michal_: i think that suse starts a lot of initscripts there, maybe look into the dev folder and find out which scripts must be disabled
1155831331 M * bonbons daniel_hozac: do you know a good way to find out what applications are dirtying inodes ervywhere? (I tried simple printk in do_fsync/do_sync but it' not that useful... especially do_sync where I just see pdflush)
1155831398 M * bonbons it's really anoying on laptop, external usb hard-disk with /home, /tmp and /var/tmp and stil get very too much writes at same time on all rw-mounted partitions on the main drive
1155831499 J * yarihm ~yarihm@vpn-global-dhcp1-46.ethz.ch
1155832836 M * michal_ coocoon: dev folder?
1155832841 M * michal_ i don't think so
1155832843 M * michal_ that's one
1155832845 M * michal_ second
1155832863 M * michal_ i have already disabled any single script except sshd,syslog,dbus
1155832890 M * michal_ and configured apps/init
1155832899 M * michal_ cmd.start,stop,runlevel
1155832907 M * michal_ everyhting possible
1155833016 M * derjohn bonbons, as you are a c coder: Could you please help me out with vfs_unlink? http://paste.linux-vserver.org/252
1155833106 M * bonbons derjohn: do you have the location where vfs_unlink() is defined/implemented available?
1155833133 M * derjohn sure ... fs/exec.c in ubuntu's kernel, line 1584
1155833139 M * derjohn i can post, too
1155833237 M * bonbons in fs/exec.c I have no occurence of vfs_unlink (2.6.17.7 + vs2.1.1rc27)
1155833240 Q * h01ger Quit: h01ger
1155833243 M * derjohn bonbons, 1593 are somewhat large to pastebin, the part you see is probably just plugged in
1155833268 P * JimmyGulp Leaving
1155833272 M * derjohn bonbons, it a ubuntu patches kernel. they have a) kernel crashdump extension and b) Unionfs 
1155833275 M * bonbons derjohn: just the implementation of vfs_unlink() should be enough
1155833289 M * derjohn both use vfs_unlink
1155833300 M * derjohn my compile runs fails :(
1155833358 Q * yarihm Quit: Leaving
1155833402 M * derjohn i think ten vfs_unlink was entended for the dentry limits, nor? so, in a crashdump case i would expect it's of now use anymore to prevent a dentry attack
1155833461 M * bonbons looking at the depth of vfs_unlink to see if use of last argument is always NULL-safe
1155833482 M * bonbons should be, if I remeber well someone was already digging in that area some time ago
1155833522 M * derjohn bonbons, you mean the unionfs-part? i heard ppl are using it with vserver 
1155833542 M * derjohn i just look if i find a unionfs patch for devel
1155833582 M * bonbons yep
1155833850 J * h01ger ~holger@socket.layer-acht.org
1155833871 M * bonbons hit leafs without seeing code broken by NULL for last parameter :)
1155834372 M * derjohn bonbons, eh what? maybe i should ask Bertl_zZ , he invented the dentry limit stuff ;)
1155834454 M * bonbons do changing vfs_unlink(a,b) to vfs_unlink(a, b, c) should always be fine, only drawback, you may castrate some functionality :)
1155834531 M * bonbons I think most if not all unionfs-patches just put NULL in the new argument and thus make bind-mount-extensions ineffective on union-fs
1155834600 M * derjohn bonbons, I think you can call me the master of castration soon ;)
1155834666 M * bonbons hehe, how many filesystems are you going to patch? Just the unionfs or more of them?
1155834795 N * Bertl_zZ Bertl
1155834800 M * Bertl evening folks!
1155834809 M * Bertl evening 'master of castration' :)
1155834894 M * mnemoc O_O
1155834916 M * derjohn <highvoiceon> hello Bertl :))
1155834916 M * Hollow hehe
1155834989 M * derjohn bonbons, i found a patch, you are right, they put in a NULL as third argument in vfs_unlink(...)
1155835017 M * Hollow Bertl: everything works fine now, and we can completely drop persistent
1155835017 J * ybanafa ~ybanafa@82.114.177.154
1155835036 M * Hollow while still doing external setup
1155835043 M * Bertl Hollow: hehe, okay, guess we will keep it around nevertheless
1155835055 M * Hollow yup, no problem with that
1155835061 M * Bertl welcome ybanafa!
1155835110 M * Hollow Bertl: what's the state for the accounting syscalls, btw? ;)
1155835119 M * Bertl Hollow: so which of the patches you suggested are required now?
1155835150 M * Hollow well, the dont-migrate-on-creat still makes sense for persistent IMO, and we need a way to get helpers exit code in vs_state_change
1155835153 M * ybanafa ll ..I'm getting this error :center:/home/ybanafa# vserver vs10 build \
1155835155 M * Hollow (no patch for this yet)
1155835155 M * ybanafa > -n vs10 \
1155835155 M * ybanafa > --hostname vs10.adenit.com \
1155835155 M * ybanafa > --interface eth0:192.168.0.10/24 \
1155835155 M * ybanafa > -m debootstrap -- -d sid
1155835155 M * ybanafa Root-directory '/vservers' does not exist or is invalid
1155835195 M * Bertl might it be that the 'Root-directory '/vservers' does not exist or is invalid'?
1155835234 M * ybanafa Bertl : how can Fix it (plz point me docs)
1155835246 M * derjohn mkdir /vservers ?
1155835250 M * Bertl hmm, I'd suggest something like that :)
1155835265 M * Hollow Bertl: and the put-admin-if-persistent-on-ctx-death patch is needed
1155835277 M * Bertl ybanafa: but, this looks like you might not have installed the tools properly ...
1155835320 M * Bertl Hollow: okay, I think we can safely add the no migrate (or at least make it an option/flag)
1155835358 M * ybanafa Bertl : where to  mkdir /vservers  .. I cannot find config files 
1155835367 M * Bertl Hollow: I'm also completely with you that we should feed back the results of the helpers somehow, but not to the syscall directly imho
1155835368 M * mnemoc o.o
1155835386 M * Bertl ybanafa: how did you install the tools (util-vserver I presume)
1155835393 M * Hollow Bertl: why so? not that i'm against it, but you do the same in vs_reboot
1155835418 M * Bertl yes, that's different, the reboot is handled inside
1155835434 M * Bertl the create helper returns directly to the syscall
1155835446 M * Bertl and the exit helper does not send the info anywhere
1155835450 M * ybanafa Bertl : center:/home/ybanafa# vserver-info
1155835451 M * ybanafa Versions:
1155835451 M * ybanafa                    Kernel: 2.6.14.3-vs2.0.1
1155835451 M * ybanafa                    VS-API: 0x00020001
1155835451 M * ybanafa              util-vserver: 0.30.210; Aug 13 2006, 20:25:01
1155835451 M * ybanafa Features:
1155835453 M * ybanafa                        CC: gcc, gcc (GCC) 3.3.5 (Debian 1:3.3.5-13)
1155835455 M * ybanafa                       CXX: g++, g++ (GCC) 3.3.5 (Debian 1:3.3.5-13)
1155835457 M * ybanafa                  CPPFLAGS: ''
1155835458 M * ybanafa                    CFLAGS: '-g -O2 -std=c99 -Wall -pedantic -W'
1155835460 M * ybanafa                  CXXFLAGS: '-g -O2 -ansi -Wall -pedantic -W -fmessage-length=0'
1155835465 M * ybanafa                build/host: i686-pc-linux-gnu/i686-pc-linux-gnu
1155835465 M * ybanafa              Use dietlibc: yes
1155835466 M * ybanafa        Build C++ programs: yes
1155835468 M * ybanafa        Build C99 programs: yes
1155835469 M * derjohn ybanafa, pastebin !!!
1155835470 M * ybanafa            Available APIs: v13,net
1155835472 M * ybanafa             ext2fs Source: e2fsprogs
1155835474 M * Bertl ybanafa: okay, first, for everything longer than 3 lines, please use paste.linux-vserver.org
1155835474 M * ybanafa     syscall(2) invocation: alternative
1155835478 M * ybanafa       vserver(2) syscall#: 273/glibc
1155835480 M * ybanafa Paths:
1155835482 M * ybanafa                    prefix: /usr/local
1155835485 M * ybanafa         sysconf-Directory: ${prefix}/etc
1155835487 M * ybanafa             cfg-Directory: ${prefix}/etc/vservers
1155835489 M * ybanafa          initrd-Directory: $(sysconfdir)/init.d
1155835489 M * mnemoc .oO
1155835491 M * ybanafa        pkgstate-Directory: ${prefix}/var/run/vservers
1155835492 M * Hollow floodserv?
1155835493 M * ybanafa           vserver-Rootdir: /vservers
1155835493 M * Bertl ybanafa: otherwise you end up on many /ignore lists :)
1155835503 M * ybanafa sorry 
1155835519 M * ybanafa this the 1st time i use irc
1155835532 M * Bertl np, you'll learn :)
1155835544 M * Hollow Bertl: so, even if the helper fails ctx_create should return 0?
1155835569 M * Bertl Hollow: no, imho we need a special return, and/or a way to return the helper codes reliable
1155835601 M * ybanafa Bertl : installed util-vserver = /usr/src/.configure ..make ..make install ..when I do testme.sh everything = working
1155835606 M * Bertl Hollow: we could for example have a field in the create call
1155835624 M * Bertl ybanafa: okay, when you did 'make isntall'
1155835633 M * Hollow yeah, why not.. i don't have a preference here
1155835640 M * Hollow as long as i can get it ;)
1155835641 M * Bertl ybanafa: it said something about a make distro-install or so
1155835660 M * Bertl ybanafa: please check the logs (or do the install again) and add this one too
1155835682 M * Bertl Hollow: yeah, will think about it and find a solution
1155835689 Q * suka Remote host closed the connection
1155835701 M * Hollow another question here is: should the context be created even if the helper failed (i.e. migrate after create + claim on persistent)
1155835703 M * Bertl Hollow: the +admin magic will have to wait a little, as we have no trigger there available
1155835718 M * ybanafa Bertl : thank I'll do reinstall
1155835825 M * Bertl ybanafa: the second install (the distro part) creates all the required stuff
1155835843 M * Bertl ybanafa: what distro do you use? there might be a distro package too
1155835846 N * lolilol olilo
1155836019 M * Hollow off to dinner, bbl
1155836056 J * gerrit_ gerrit@12.174.94.104
1155836065 J * gerrit__ gerrit@1010ahost104.starwoodbroadband.com
1155836073 Q * gerrit_ 
1155836129 M * ybanafa Bertl : I reinstalled util again & working ..thanks :)
1155836188 M * ybanafa Bertl : thank You ! 4 this great VPS
1155836238 M * michal_ Bertl: howk :)
1155836259 M * michal_ Bertl: ever tried to put suse into vserver? or do you know about some doc describing it?
1155836275 M * michal_ after a lot of failed vserver suse start
1155836279 M * Bertl ybanafa: you're welcome! have fun!
1155836302 M * michal_ decided to write my own /etc/init.d/rc script - and did one - haev putonly mount -a and /etc/init.d/sshd start there :)
1155836323 M * Bertl michal_: I don't see a problem with suse when you use the sysv init mode
1155836325 M * michal_ and now, i can see running vserver in vserver-stat
1155836353 M * michal_ but cannot enter it. but can say, chctx to suse's one and processes are visible...etc
1155836354 M * Bertl michal_: might be that you need to adjust the rc script for the plain init style
1155836371 M * michal_ ok...how do i chose sysv style?
1155836393 M * Bertl echo "sysv" > /etc/vservers/<name>/apps/init/style
1155836417 M * michal_ vserver suse stop -> vserver 'suse' is not running
1155836418 M * michal_ ups
1155836429 M * Bertl try vkill
1155836438 M * michal_ it is...chcontext --ctx 2 ps aux shows me processes...
1155836550 M * michal_ vps aux | grep syslog shows me pid 2105 in ctx 2...
1155836564 M * michal_ vkill --xid 2 -- 2105
1155836575 M * michal_ vc_ctx_kill() no such process
1155836617 M * Bertl that's interesting ... Hollow, you ahd similar with xid=2 or with other xids too?
1155836628 M * Bertl michal_: what kernel patches?
1155836643 M * Bertl michal_: does /proc/virtual list the '2'?
1155836663 M * michal_ plain vserver + 2.6.16.27
1155836667 M * michal_ nothing else (yet)
1155836676 M * michal_ vs2.1.1-rc22
1155836676 M * Bertl 2.6.16.27?
1155836688 M * michal_ yes, patched vserver with .27
1155836698 M * Bertl what's that?
1155836710 M * michal_ vanilla kernel version :)
1155836721 M * Bertl they are at .27 now?
1155836727 M * michal_ mhm
1155836732 M * michal_ yes, it sounds strange ;)
1155836760 M * michal_ anyway, had no prob...until i've decided to enclose suse ;)
1155836779 M * Bertl well, it might trigger a bug somehow .. please check the proc
1155836793 M * Bertl have to leave for a few minutes, but will be back shortly
1155836801 N * Bertl Bertl_oO
1155836820 M * michal_ it does
1155836918 Q * olilo Remote host closed the connection
1155837203 M * michal_ ok...removed mine rc..got that one from suse back in the place. sysv initstyle chosen.
1155837226 M * michal_ vshelper.init can not determine xid of vserver 'suse' returned value was ''
1155837251 J * olilo hiddenserv@tor.noreply.org
1155837326 M * Hollow Bertl_oO: yep, but i wasn't able to reproduce it yesterday anymore.. seems to happen quite randomly..
1155837403 J * ola ~ola@host92-107.pool870.interbusiness.it
1155837408 M * ola hello
1155837422 M * ola I'm trying to install vanilla kernel on ubuntu dapper with vserver 
1155837451 M * michal_ hey ola
1155837473 M * ola I've compiled with initrd and even compiled statically ext3 support but at boot it fails with kernel panic telling me that it doesn't know how to read fs
1155837480 M * ola what's wrong?
1155837504 Q * ybanafa Quit: Leaving
1155837507 M * michal_ Bertl_oO: got a phone call...will be back a bit later(max 2h). i hope you will be tonight online ;)
1155837685 M * daniel_hozac ola: without the error message, and a description of your setup, hard to tell.
1155837752 M * ola daniel_hozac, u are right :) I'll reboot and I'l wrote it
1155838109 M * pisco hi, had someone some experience with the openqrm management tool an vservers?
1155838253 Q * cehteh Ping timeout: 480 seconds
1155838353 M * pisco looks very compromising to have different virtualisation technologies under one hat
1155838478 M * s0undt3ch hello ppl
1155838497 M * s0undt3ch I have a guest /tmp with 16 megs, can I maje it bigger?
1155838519 M * Hollow s0undt3ch: /etc/vservers/<name>/fstab
1155838531 M * Hollow did you hit the limit with clamav? :p
1155838539 M * s0undt3ch aparently yes
1155838551 Q * shedi Quit: Leaving
1155838551 M * s0undt3ch you're whatching me somewere?
1155838564 M * Hollow no, i had the same problem yesterday
1155838577 M * s0undt3ch lol
1155838583 M * Hollow the new database seems >=16M
1155838687 M * s0undt3ch yeah, well, at least it forced me to run clamav as qscand user again, I was running it as roo because of some troubles in the past, that aparently don't occur anymore :)
1155838770 M * bonbons Hollow, s0undt3ch: weird, I saw just the same, yesterday as well ;)
1155838787 M * Hollow heh
1155838805 M * s0undt3ch ppl, vserver defaults must be raised just because of clamav :=
1155838819 Q * ola Quit: Leaving
1155838906 M * daniel_hozac hehe.
1155838911 M * daniel_hozac enrico probably will.
1155838922 M * daniel_hozac (he maintains clamav in Fedora Extras)
1155838950 J * krtek ~root@agepmac3.epfl.ch
1155838960 M * mnemoc 16M is ok as default imo, even if clamav doesn't like it
1155838966 M * krtek Hi everybody
1155838975 M * s0undt3ch mnemoc: I was just kiding
1155838978 M * mnemoc =)
1155838990 M * daniel_hozac i'm fairly serious. i expect the limit to be raised :)
1155838994 M * daniel_hozac krtek: you shouldn't IRC as root :)
1155839016 M * krtek I know ^^ but it's not my first problem at this time :D
1155839050 M * s0undt3ch krtek: at least have mindent scramble that for you :)
1155839146 M * krtek I've some problem with chbind :( it says me that my kernel don't provide network virtualization, but in my config all seems to be fine
1155839152 N * Bertl_oO Bertl
1155839154 M * Bertl ack now
1155839157 M * Bertl +b
1155839177 M * Bertl krtek: sure you bootet the very same kernel?
1155839189 M * daniel_hozac krtek: except you enabled the "Show legacy version ID".
1155839211 M * krtek Bertl: Yes it's the good kernel
1155839220 M * krtek daniel_hozac: Yes, it's enabled
1155839225 M * Bertl s0undt3ch: what's the clamav issue?
1155839230 M * daniel_hozac krtek: hence causing your problem.
1155839237 M * daniel_hozac don't people read the help texts anymore?
1155839241 M * s0undt3ch Bertl: gone now, /tmp was too small
1155839264 M * Bertl s0undt3ch: ah, k
1155839280 M * daniel_hozac krtek: why did you enable it?
1155839295 M * krtek i've followed a tutorial about installing vserver on debian, and no mention about this problem... And google don't gives the answer either ^^ but it's true that I haven't read the help files 
1155839309 M * krtek it's enabled by default in the debian vserver patch. 
1155839317 M * daniel_hozac really? are you serious?
1155839354 A * daniel_hozac pokes waldi and/or micah.
1155839359 M * krtek yes, patching the kernel witch --append-patch vserver enable this option
1155839367 M * waldi daniel_hozac: ?
1155839378 M * Bertl hmm, maybe we have it on by default in stable?
1155839382 M * daniel_hozac no, i checked.
1155839390 M * micah ... humm?
1155839391 M * krtek I'll give a new try without it
1155839392 M * daniel_hozac i checked yesterday when someone else had the same problem.
1155839407 M * krtek I'm running etch for information
1155839409 M * Bertl ah, so that's the problem of the next few weeks then :)
1155839417 M * Bertl (good to know :)
1155839438 M * micah whats the issue?
1155839452 M * daniel_hozac waldi, micah: is CONFIG_VSERVER_LEGACY_VERSION enabled by default?
1155839470 M * micah # CONFIG_VSERVER_LEGACY_VERSION is not set
1155839493 M * micah at least in 2.6.16-2-vserver-686, which I would assume is the same for the others
1155839554 M * krtek I haven't take the image, I've used the package of vserver patch and then compile witch make-kpkg --append-patch vserver, and the option is enabled 
1155839587 M * micah krtek: when you apply the patch you decide what the kernel config options are
1155839607 M * Bertl maybe debian make-kpgp enables all by default (debian style?) please ignore if inappropriate :)
1155839633 M * Bertl like make allyesconfig ...
1155839672 M * krtek micah: I haven't get any question, but maybe a strange configuration, I haven't installed Debian myself
1155839719 M * micah Bertl: you have to use a .config file of your own and then it will do just like makeoldconfig does
1155839734 M * micah krtek: where did you get your kernel config file?
1155839776 M * krtek micah: I do it myself because I've to enable some module for the raid arrays
1155839833 M * micah krtek: when it applies the patch, it will ask you about the vserver options because they will not already be in your .config
1155839926 M * krtek micah: maybee I've answered the question the first time without looking at them, and when I recompiled after getting a problem, vserver options where in .config, so no questions anymore 
1155839958 M * krtek compilation completed, I reboot and give a test, see you in some minutes
1155839990 Q * krtek Quit: Lost terminal
1155840139 M * Bertl okay, so just for the record, debian has sane defaults, and the legacy version is off by default
1155840217 J * krtek ~krtek@agepmac3.epfl.ch
1155840365 Q * krtek 
1155840526 J * shedi ~siggi@inferno.lhi.is
1155840664 Q * s0undt3ch Ping timeout: 480 seconds
1155840762 J * krtek ~krtek@agepmac3.epfl.ch
1155840770 M * Bertl wb krtek!
1155840793 M * krtek Hi again, all's fine now, thx a lot :)
1155840803 M * Bertl Hollow: a version 2 create with return values from the helper would be fine?
1155840815 M * Hollow sure
1155840897 M * Bertl okay, now what return code should we send if the helper fails?
1155840908 M * Hollow -1 + errno?
1155840914 M * Bertl and shall we continue with the context creation?
1155840919 M * Hollow imo no
1155840929 A * mnemoc nods 
1155840937 M * Bertl -1 == EPERM, no?
1155840950 M * Bertl and what does that mean if the helper isn't present
1155840954 M * mnemoc errno != return value :p
1155840961 M * Hollow :9
1155840963 M * Hollow :)
1155840966 M * Bertl mnemoc: not for the kernel :)
1155840969 M * mnemoc :(
1155840973 M * Hollow yeah, i was talking about userspace
1155840983 M * Bertl that's your part *G*
1155840992 M * mnemoc ^_^
1155840996 M * Hollow well, if sc_helper is enabled and vshelper is not configured correctly, we should fail too imo
1155841076 J * s0undt3ch hmcwdlqm@bl8-1-150.dsl.telepac.pt
1155841121 M * Bertl I would be inclined to abuse ENOEXEC for error cases
1155841134 M * Hollow sounds ok to me
1155841136 M * Bertl (in which case we _have_ to abort context creation)
1155841186 M * Bertl alternatively, we could 'just' pass the script return code
1155841204 M * Bertl and let userspace take care of the rest, note: this could be advanageous
1155841218 M * Bertl i.e. you could also return non-letal info from the helper
1155841253 Q * lilalinux Remote host closed the connection
1155841262 M * daniel_hozac but then shouldn't the rest of vc_ctx_create be executed as well?
1155841268 M * Hollow would also mean we need to kill the context if sth failed, no?
1155841296 M * daniel_hozac i think we should just make it fail if the helper fails.
1155841303 A * Hollow nods
1155841347 M * daniel_hozac if it's not fatal, let the helper log it and return 0 anyway.
1155841384 M * Hollow that's what we do currently... most things are logged and either EXIT_SUCCESS or EXIT_FAILURE is returned
1155841446 M * Bertl what about the following compromise ...
1155841470 M * Bertl error code <= 0 means termination, positive values are return codes
1155841483 M * daniel_hozac can you do negative exit codes?
1155841490 M * daniel_hozac oh, you mean for the syscall, nevermind.
1155841516 M * Bertl hmm, I mean from the helper
1155841557 M * Hollow can't you return just 0-255?
1155841568 M * daniel_hozac yeah, that's what i thought.
1155841574 M * daniel_hozac well, 127 actually.
1155841634 M * Hollow even less..
1155841644 M * Bertl okay, so we could still shift that range and/or designate special codes?
1155841656 M * Hollow well, sounds quite hackish
1155841723 M * daniel_hozac yeah. do we even need to forward the helper's return code? wouldn't logging and a special "helper failed" return code (like ENOEXEC) work?
1155841754 M * Hollow yeah, i'd prefer a solution like this
1155841828 M * Bertl okay, so we do not want to handle the helper return code in any case, that's perfectly fine with me :)
1155841853 M * Hollow yup, we just need success or failure
1155841861 M * Hollow and react properly
1155841910 M * Hollow just like in vx_reboot ;)
1155841917 M * Hollow *vs_reboot
1155842053 Q * zkbrsnie Read error: Connection reset by peer
1155842065 M * daniel_hozac Hollow: btw, did you get my questions the other day? about clone and libvserver compatibility stuff?
1155842075 M * Hollow no?
1155842116 M * Hollow i recently changed sys_clone back to _syscall2 instead of _syscall5
1155842128 M * Hollow i don't know why, but it fixed some problems
1155842131 M * daniel_hozac 2006-08-16T00:49:18 < daniel_hozac> Hollow: i guess you won't be having compatibility stuff in libvserver?
1155842137 M * daniel_hozac 2006-08-16T00:55:31 < daniel_hozac> Hollow: and does _syscall5 avoid all of the sparc/s390 oddities?
1155842167 M * Hollow ah, do you know about syscall5 issues?
1155842183 M * daniel_hozac no, but we had to fix util-vserver for sparc and s390.
1155842216 M * daniel_hozac sparc returns the parent's pid in the child.
1155842220 M * Bertl okay, the return code warning is already there in do_vshelper() I assume that is sufficient then
1155842222 M * daniel_hozac s390 has the arguments in the other order.
1155842242 M * daniel_hozac s/other/reversed/
1155842268 M * Hollow i see... will take a look at util-vserver clone stuff
1155842275 M * daniel_hozac http://daniel.hozac.com/vserver/util-vserver/util-vserver-0.30.210-clone-arch.patch
1155842307 M * Hollow ah yes... i even saw that patch already
1155842310 M * mnemoc Hollow: there is more debug info on vcd than what you get running it with -d ?
1155842340 M * Hollow mnemoc: you can pass XMLRPC_TRACE_XML=1 in the environment, and xmlrpc-c library will print all raw xml data
1155842376 M * mnemoc thanks
1155842385 M * Hollow but else you're pretty much left with log_debug() ;)
1155842399 M * Hollow or strace, gdb if you need to track special bugs
1155842462 M * Hollow daniel_hozac, Bertl: with syscall5 i got the "Bad address" error like the one we already investigated in exec-ulimit...
1155842478 M * Hollow but maybe that was a fault on my side..
1155842639 M * mnemoc Hollow: do you know any trick to solve "[error] cannot listen: Address already in use" after vcd was killed?
1155842663 M * Hollow wait :)
1155842676 M * mnemoc ok :)
1155842690 M * Hollow will work again quite fast.. 30 secs or so
1155842704 A * mnemoc counting
1155842713 M * Hollow i already fixed most of the already in use issues, but it sometimes still happens
1155842726 M * mnemoc go :)
1155842754 Q * krtek Quit: leaving
1155843064 M * Bertl doener, daniel_hozac, Hollow: what about this (only for 'normal' contexts for now)
1155843068 M * Bertl http://vserver.13thfloor.at/Experimental/delta-helper-feat01.diff
1155843153 M * Hollow shouldn't out_unhash fail with ENOEXEC?
1155843186 M * Hollow also persistent has to be moved below vs_state_chajnge, no?
1155843221 Q * ntrs_ Read error: Connection reset by peer
1155843229 J * ntrs ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com
1155843234 M * Hollow also the same needs to be changed in network.c
1155843285 M * Hollow ah, you probably meant that with "normal" contexts
1155843352 Q * hallyn Quit: leaving
1155843448 M * Bertl Hollow: ah, good point, the return code, well, only in the helper case
1155843522 J * Wenix ~wenix@81.7.189.11
1155843631 M * Bertl wb Wenix!
1155843753 Q * coocoon Quit: KVIrc 3.2.0 'Realia'
1155843777 M * Bertl okay, http://vserver.13thfloor.at/Experimental/delta-helper-feat02.diff
1155843810 M * Hollow yep, looks better... what about the persistent thing?
1155843852 M * Bertl will address that now, but we should test this with error pathes (failure case and such) regarding the unhash
1155843864 M * Bertl (that's why I did that one first)
1155843903 J * coocoon ~coocoon@p54A0770C.dip.t-dialin.net
1155843921 M * Hollow ok, will test it..
1155843944 M * michal_ Bertl: ping?
1155843948 M * michal_ root
1155843950 M * michal_ y
1155843952 M * michal_ not here :P
1155843971 M * Bertl pong!
1155844009 M * michal_ so, ssyv init style, original rc from suse. vserver set to boot init level 5
1155844081 M * michal_ can not determine xid of vserver...
1155844097 M * Hollow Bertl: what will happen if persistent is enabled and you unhash the context?
1155844103 M * Bertl michal_: did you set one?
1155844115 M * michal_ sure
1155844132 M * michal_ vservers/suse/context?
1155844135 M * michal_ set to 2
1155844144 J * ola aa@host206-108.pool8254.interbusiness.it
1155844149 M * ola back
1155844150 M * Bertl wb ola!
1155844173 M * Bertl michal_: do you have debugging enabled in your kernel?
1155844177 Q * gerrit Quit: KVIrc 3.2.0 'Realia'
1155844198 M * Bertl michal_: would be cool to get some debug outputs
1155844199 M * ola well I take note of the error booting vserver patched kernel on a laptop 'kernel panic not syncing VFS unable to mount root fs on unknown_blok (0,0)
1155844202 M * ola blok
1155844223 M * Bertl looks like your filesystem is not supported in the kernel, or the device is not present
1155844225 M * sid3windr heya Bertl 
1155844234 M * Bertl ola: i.e. device driver missing
1155844256 M * ola uhm it's a notebook
1155844262 M * michal_ Bertl: actualy...let's make another try - just found something interesting on a mailing list - seems like suse rc script won't work with vserver...i'm trying with provided by that guy now...
1155844263 M * ola so I need some special driver for my disk?
1155844266 M * michal_ will see
1155844276 M * michal_ and that problem with unable to kill? already restarted machine :]
1155844316 M * phreak`` ola: sure, either compiled into the kernel or as module .. like fs-driver (e.g. ext2/ext3/xfs/reiserfs/..)
1155844328 M * ola uh ultra ata storage controller
1155844332 M * mnemoc yuck yuck, vcd got blocked again :\
1155844336 M * ola i got fs driver
1155844350 M * Bertl michal_: no idea, any input is appreciated
1155844352 M * ola maybe I miss ultrata
1155844355 M * ola c u later
1155844358 Q * ola Read error: Connection reset by peer
1155844369 M * phreak`` Bertl: great to see some of your work getting into upstream (the -ro bind mounts)
1155844370 M * Bertl michal_: but in general I'd suggest to try with 2.6.17.8-rc29 (just to make sure)
1155844401 M * Hollow BME in mainline?
1155844401 M * Bertl phreak``: well, yeah, since the IBM folks spent endless hours to change it over and over
1155844403 M * Hollow wh00t!
1155844435 M * phreak`` Hollow: you should really read lkml from time to time :)
1155844437 M * Bertl phreak``: I'm pretty sure I will not even be mentioned, but actually I don't care much, the functionality will be there
1155844473 M * Hollow phreak``: well, i do... but only once a week or so
1155844486 M * sid3windr :)
1155844493 M * phreak`` Bertl: well I know that it originated from your work, so I see it as _your_ work (ok, not only) getting into upstream
1155844646 M * sid3windr :)
1155844661 A * sid3windr should stop smiling ;)
1155844671 M * Hollow too much drugs? ;)
1155844677 M * Bertl phreak``: well, Dave is doing a very good, very patient job here
1155844691 M * Bertl phreak``: I would have given up a long time ago :)
1155844704 M * Hollow reminds me of baselayout-vserver
1155844710 M * Hollow i gave up a long time ago :p
1155844712 M * phreak`` oO
1155844719 A * phreak`` runs
1155844737 M * sid3windr Hollow: nah, not really, just not a lot to say ;)
1155844743 M * sid3windr but still wanted to acknowledge :p
1155844751 M * Hollow although i think with VXC_SECURE_MOUNT we could make baselayout workable..
1155844758 M * Hollow s/workable/accepted/
1155844760 M * Hollow ;)
1155844764 M * phreak`` heh
1155844769 A * sid3windr ponders
1155844780 M * sid3windr Bertl: did you check my oops yet? do you need more info about it?
1155844786 M * Hollow the only thing i worry about is sulogin in /sbin/rc
1155844802 M * Hollow because we have no /dev/console and a reboot -f afterwards might result in a loop
1155844809 J * mkhl mkhl@200-153-153-102.dsl.telesp.net.br
1155844824 M * phreak`` Hollow: btw, I added the patches of Bertl's (2 iirc) and daniel_hozac (1 iirc) to the util-vserver patch repo
1155844832 M * Hollow yep, i saw that
1155844840 M * Hollow but you missed the clone-arch patch from daniel
1155844840 M * michal_ ok
1155844843 M * michal_ suse won't work
1155844854 M * michal_ it's rc script can be just trashed
1155844867 M * michal_ and replacing it with another one is not a smart idea anyway
1155844922 M * coocoon michal_: do u have another suse dists which will work? 
1155844938 M * michal_ y?
1155844955 M * michal_ suse won't work inside vserver. and that's suse fault.
1155844958 M * michal_ that's all about it :)
1155844964 M * coocoon only wanted to know
1155845038 M * Bertl sid3windr: no, completely missed it, url?
1155845057 M * sid3windr (sent it to the ml) http://magic.powersource.cx/~tom/vserver-oops.pn
1155845058 M * sid3windr g
1155845058 M * Hollow michal_: maybe http://lylix.net/vps+templates/func,select/id,1/orderby,2/page,2/ helps?
1155845066 M * sid3windr lost the g there ;)
1155845078 M * sid3windr but I'm afraid it's not enough and you need the lines above it :/
1155845145 J * ola ~ola@host79-128.pool8251.interbusiness.it
1155845181 M * ola well I suppose the <*> ATA/ATAPI/MFM/RLL support is enough :/
1155845187 M * michal_ that's a very tupid addr if you have to write it manualy :P
1155845202 M * ola [17179574.172000] Probing IDE interface ide0...
1155845202 M * ola [17179574.460000] hda: IC25N080ATMR04-0, ATA DISK drive
1155845242 M * michal_ Hollow: not even a option for me. this are full images to be downloaded.
1155845250 M * michal_ nd i have to run suse i have on disk
1155845258 M * Hollow oh, i see..
1155845271 M * michal_ downloading this suse would take few days...than another weeks for packages:)
1155845309 Q * Viper0482 Remote host closed the connection
1155845313 M * michal_ if you have a great connnection you could look into that opensuse image to see if its rc script is any different ;)
1155845329 M * coocoon michal_: maybe i can start a guest on my server and gave u access to it than u can hava a look for it how to get it to work
1155845332 M * ola I've to use PCMCIA to notebook disks?
1155845342 M * coocoon a suse guest i mean
1155845355 M * michal_ coocoon: sure, it'd like to.
1155845363 M * michal_ which version is it?
1155845376 M * michal_ have you been replacing susestupid rc script?
1155845378 M * Bertl sid3windr: do you have the kernel at hand? i.e. could you do a few address to line conversions for me?
1155845394 M * coocoon which u want, the only one which i am not having is 10.1
1155845395 M * sid3windr I guess yes
1155845399 M * mnemoc [Thu Aug 17 22:05:55 2006] [ 1722] [warn] Segmentation fault *G*
1155845402 M * sid3windr I don't know how but if you teach me ;)
1155845419 M * Bertl sid3windr: sure, np, first let's try with addr2line
1155845420 M * michal_ coocoon: newest one possible :)
1155845444 M * coocoon hm have tried to install it via yum and apt but it failled, but u can update it if u want
1155845452 M * Hollow mnemoc: can you paste some [trace] lines before?
1155845454 M * coocoon hehe but if it then work i do not know
1155845454 M * Bertl sid3windr: like this: 'addr2line -e vmlinux c01625da
1155845470 M * coocoon i think it will not but maybe give it a try
1155845516 M * sid3windr pixel:/usr/src/linux-2.6.17.7-vs2.0.2-rc28# addr2line -e vmlinux c01625da
1155845516 M * sid3windr ??:0
1155845518 M * sid3windr funky ;
1155845519 M * sid3windr ;)
1155845534 M * coocoon michal_: which one u need 10.0 x86 or x86_64
1155845551 M * sid3windr Bertl: should we continue this in private perhaps? :)
1155845575 M * michal_ coocoon: x86 plase
1155845579 M * michal_ *please
1155845584 M * coocoon ok wait a moment
1155845602 M * Bertl sid3windr: no, it's fine here, so your addr2line is broken (well most are, they are missing an important patch)
1155845613 M * Bertl sid3windr: np, we use gdb for that
1155845619 M * sid3windr heh
1155845682 P * stefani I'm Parting (the water)
1155845715 M * Bertl sid3windr: 'gdb vmlinux' then l *0xc01625da
1155845743 M * sid3windr No symbol table is loaded.  Use the "file" command.
1155845776 M * Bertl hmm, so your kernel is missing the debug symbols then
1155845778 M * michal_ coocoon: and also tell me what have you put into vservers/suse or whatever is the name ;)
1155845796 M * sid3windr apparently
1155845809 M * sid3windr should I enable, recompile, reboot, make new screenshot, reboot, gdb again? :)
1155845825 M * Bertl question is, is that reproducible?
1155845833 M * sid3windr it happens every time
1155845837 Q * ola Remote host closed the connection
1155845845 M * Bertl excellent, then please let's enable a few other options too
1155845848 M * coocoon micah: dunno what u mean
1155845859 M * sid3windr okay, tell me which :) (and where I can enable the debugging symbols)
1155845865 M * Bertl sid3windr: could you upload the current kernel config somewhere?
1155845872 M * coocoon michal_: sorry u were meant ;-)
1155845873 M * sid3windr oki
1155845889 M * michal_ coocoon: i mean it would be good to know what have you put into <vserver configuration dir>/apps/init :)
1155845896 M * michal_ for that suse
1155845899 M * sid3windr Bertl: http://magic.powersource.cx/~tom/vserver-config
1155845914 M * micah coocoon: ?
1155845924 M * micah coocoon: ah, wrong nick
1155845960 M * coocoon micah: sorry
1155845966 M * Bertl sid3windr: okay, take the config and remove the following lines with an editor
1155845973 M * Bertl # CONFIG_UNWIND_INFO is not set
1155846034 J * Savvy ~SavvyChat@81.215.86.76
1155846046 M * Bertl sid3windr: # CONFIG_DEBUG_KERNEL is not set
1155846050 M * coocoon michal_: be honest nothing hehe have create them with vserver foo built -yum .... or apt-rpm
1155846063 M * coocoon +d
1155846100 M * Bertl sid3windr: # CONFIG_VSERVER_DEBUG is not set
1155846106 M * coocoon michal_: 90 mb is too many for u to download
1155846125 M * michal_ yeppp
1155846130 M * Bertl sid3windr: then do 'make oldconfig' and enable the various debug options, including vserver debug and history
1155846133 M * michal_ well - i just need some informations
1155846136 M * coocoon ok wiat a moment i will hope to get it to work
1155846150 M * michal_ suse guest configuration & init suse configurtions
1155846151 M * michal_ that's all :)
1155846175 M * michal_ i'm on gprs currently (one week more actualy, than back on ful speed)
1155846214 M * sid3windr compiling
1155846241 M * Bertl sid3windr: please upload the config, so that I can double check
1155846259 M * matti Bertl, michal_ :)
1155846268 P * pisco 
1155846287 M * Bertl evening matti!
1155846295 M * sid3windr oki
1155846332 M * sid3windr Bertl: http://magic.powersource.cx/~tom/vserver-config-debug
1155846347 Q * Hollow Quit: Konversation terminated!
1155846434 M * Bertl sid3windr: ah, well, you probably want to disable the RCU torture tests :)
1155846447 M * sid3windr o :)
1155846463 M * sid3windr anything else? :)
1155846483 M * michal_ hey matti :]
1155846492 J * Hollow ~hollow@2001:a60:f026::1
1155846492 M * Bertl we probably dont want the FORCED_INLINING
1155846511 M * Bertl sid3windr: and the PAGEALLOC can go too
1155846523 M * Bertl the rest should not hurt
1155846525 M * sid3windr okay
1155846535 M * sid3windr compiling some more :)
1155846773 M * michal_ rcu torture tests _are_ tortures 
1155846781 M * michal_ have once loaded module
1155846796 M * michal_ and it was really hard to get control back over machine :P
1155846809 M * michal_ with very high load,cpu usage...etc
1155846936 M * michal_ matti: how's life? and your cat?
1155847033 M * matti michal_: Well, depends on point of view.
1155847062 M * michal_ like everything :]
1155847096 M * Hollow Bertl: setting kernel.vshelper to a non-exisiting file, i get vxD: /foo: (startup 42) returned sync with 0
1155847157 M * sid3windr (building modules..)
1155847186 M * Hollow Bertl: using an existing helper which fails, i get an oops
1155847206 M * Hollow http://paste.linux-vserver.org/256
1155847213 M * Bertl give me a few minutes, I'm currently cleaning up the stuff
1155847231 M * Hollow sure, no panic ;)
1155847246 M * matti Hehehe.
1155847248 M * matti Hollow: :)
1155847253 M * michal_ ;p
1155847262 M * Hollow huhu matti
1155847280 M * Bertl Hollow: nah, that's written on my guide, in big friendly letters :)
1155847288 A * Hollow wonders why matti is always smiling ;)
1155847306 M * michal_ dd if=/dev/urandom of=/dev/matti/mind/behaviour
1155847314 M * Hollow :)
1155847316 M * michal_ bs=1k count=1
1155847318 M * michal_ :P
1155847429 M * Bertl Hollow: okay, here is my suggestion:
1155847431 M * Bertl http://vserver.13thfloor.at/Experimental/delta-persist-fix01.diff
1155847449 M * Bertl http://vserver.13thfloor.at/Experimental/delta-helper-feat03.diff
1155847530 M * Hollow yep, looks good
1155847543 M * matti michal_: ;-p
1155847605 M * matti Hollow: I am trying to keep positive thinking :-p
1155847635 M * Hollow ok, i do as well: :)
1155847641 M * Bertl matti: and this is the right way to go ...
1155847677 M * Bertl http://www.thebards.net/music/lyrics/Always_Look_Bright_Side_Life.shtml
1155847781 M * sid3windr woops
1155847783 M * sid3windr Bertl: rebooting :)
1155847817 M * matti Bertl: :-)
1155847835 M * sid3windr haha
1155847835 Q * mkhl 
1155847837 M * sid3windr Bertl: now it works.
1155847848 M * sid3windr Bertl: must be some option that got toggled
1155847961 M * daniel_hozac sid3windr: that's the way it goes :) enable debugging and the bugs vanish...
1155847966 M * Bertl sid3windr: well, you know bisection?
1155847989 M * sid3windr ehh
1155847992 M * sid3windr I think so?
1155847993 M * sid3windr :)
1155848097 Q * bonbons Quit: Leaving
1155848208 M * michal_ Bertl: u know - it looks strange
1155848227 M * michal_ vserver suse start
1155848232 M * michal_ nothing in return
1155848240 M * michal_ vserver-stat shows running vserver
1155848246 M * michal_ which is running/not running :)
1155848252 M * matti michal_: SuSE is very shy.
1155848258 M * michal_ :P
1155848260 M * matti ;
1155848263 M * michal_ i like it 
1155848272 M * Bertl michal_: well, it is kind of 'normal' that not much is shown (with certain init styles)
1155848289 M * Bertl michal_: the question now is, what is the running/not running part
1155848290 M * michal_ but i cannot enter it later ;)
1155848308 M * Bertl michal_: you are using what patches right now?
1155848323 M * michal_ vserver suse enter -> vserver...suexec is suposed for running vservers only; aborting
1155848343 M * michal_ like i've told - 2.6.16.27-vs2.1.1-rc2
1155848362 M * Bertl hmm, what about: 'vserver suse exec bash
1155848366 M * michal_ 2.6.16 because i'll be merging it with xen later...
1155848384 M * Wonka xen is a pain with 2.6.16.27...
1155848387 M * michal_ same error cmd as prevoiusly
1155848402 M * Wonka at least, patching the patches into 2.6.16.27 is
1155848403 Q * dna Quit: Verlassend
1155848410 M * Bertl michal_: and please try with 2.6.17.8-r29 (just to make sure that the error persists)
1155848432 M * michal_ ok
1155848434 M * michal_ btw
1155848445 J * suka ~suka@bi01p1.co.us.ibm.com
1155848454 M * michal_ chcontext --xid 2 -- /bin/bash && ps aux when inside shows me
1155848485 M * michal_ faked init, some cmd that has been run from 'local' initscript and just bash
1155848494 M * michal_ so none of services has been run
1155848524 M * Bertl what state is that 'local' script in?
1155848534 Q * suka 
1155848547 M * Bertl I presume the init scripts 'hang' somewhere, and the context start did not complete for whatever reason
1155848555 M * michal_ sleeping
1155848558 M * michal_ but otherwise working
1155848659 M * michal_ this local is etc/init.d/boot.local, where i can put mine own cmds...
1155848689 M * michal_ would vserver --debug suse start tell you something?
1155848830 M * Bertl it might, i.e. it's worth a try, but _please_ let's try with an official kernel first :)
1155848849 M * michal_ sure
1155848865 M * Bertl it wouldn't be the first time that subtle changes in the kernels break something
1155848886 M * michal_ heh, tell me about it - seen it too many times...
1155849217 M * sid3windr Bertl: I got it to bork again, but differently
1155849228 M * Bertl okay, let's upload that one
1155849275 M * sid3windr http://magic.powersource.cx/~tom/vserver-crash-2.png
1155849283 M * sid3windr rebooting to working kernel again :)
1155849393 M * sid3windr http://magic.powersource.cx/~tom/vserver-config-3
1155849607 M * daniel_hozac michal_: why can't you use Xen on 2.6.17?
1155849725 M * Bertl sid3windr: hmm, there is probably a completely different oops right 'above' that
1155849734 J * DreamerC_ ~dreamerc@59-112-7-56.dynamic.hinet.net
1155849744 M * Bertl sid3windr: it would be cool to get the entire boot log ... maybe via serial?
1155849776 M * sid3windr welp
1155849779 M * sid3windr no can do :/
1155849788 M * sid3windr it's someone else's vmware server and I don't have access to that
1155849798 M * sid3windr I can try to increase to 80x50
1155849801 M * sid3windr or higher
1155849806 M * Bertl well, maybe it can do console logging?
1155849824 M * Bertl and if not, maybe switch to something better, like QEMU?
1155849837 M * sid3windr don't have access to that either ;)
1155849867 M * sid3windr booting to 80x60
1155849884 M * sid3windr eck
1155849891 M * sid3windr some init script is resetting to 80x25
1155850082 M * michal_ Bertl: btw - this unofficial patch for vserver was working well for over a week with few debina guests :]
1155850089 M * michal_ *debian
1155850097 M * michal_ downloading official one now...
1155850126 M * Bertl sid3windr: well, maybe give QEMU a try :)
1155850138 M * sid3windr I only have an account through the vmware console
1155850143 Q * DreamerC Ping timeout: 480 seconds
1155850145 Q * Savvy Quit: Download IceChat at www.icechat.net
1155850146 M * sid3windr no system account
1155850148 M * sid3windr so no qemu
1155850162 M * sid3windr I can add a serial port which dumps to a file
1155850166 M * sid3windr but I can't reach that file ;/
1155850248 A * sid3windr continues to try ;)
1155850313 M * michal_ daniel_hozac: is thre xen for .17 already avaiable?
1155850343 M * daniel_hozac michal_: i hope so, Fedora has been shipping .17 for months with Xen... ;)
1155850386 M * michal_ half-official vendor hacks are barely interesting:)
1155850590 M * sid3windr Bertl: http://magic.powersource.cx/~tom/vserver-crash-3.png
1155850783 M * Bertl hmm, interesting results:
1155850787 M * Bertl         /* This kernel should never had been booted. */
1155850787 M * Bertl         BUG_ON(!kallsyms_addresses);
1155850811 M * Bertl I'm supsecting your gcc to do strange things
1155850826 M * sid3windr :/
1155850830 M * michal_ gentoo?;p
1155850835 M * sid3windr I just notice that it now oopses when starting vmware stuff
1155850836 M * Bertl let's try a 'make distclean' (save the config)
1155850838 M * sid3windr I will remove that and try again
1155850841 M * sid3windr michal_: debian
1155850847 J * Aiken ~james@tooax7-052.dialup.optusnet.com.au
1155850880 M * Bertl welcome Aiken!
1155850921 M * sid3windr boots fine now
1155850931 M * sid3windr so I haven't found yet what broke it earlier on
1155850989 M * Aiken hello Bertl
1155851402 M * michal_ time make -j4 && let's wait ;)
1155851424 M * michal_ actualy, it's going to be the active waiting
1155851444 M * michal_ a lot of things to break....y....i mean work on :P
1155851536 Q * transacid Quit: leaving
1155851662 J * suka ~suka@bi01p1.co.us.ibm.com
1155851677 M * Bertl welcome suka!
1155851704 M * suka Hi Hebert
1155851738 M * doener morning folks
1155851772 M * Bertl good morning doener!
1155851786 M * daniel_hozac morning ;)
1155851972 M * Bertl Hollow: does the latest version work for you? i.e. with helper returning error and such?
1155852056 J * cehteh ~ct@cehteh.homeunix.org
1155852077 M * Hollow no, as said before, if i configure a non-existant helper, i get "vxD: /foo: (startup 42) returned sync with 0" with a valid helper configured which returns!=0 i get a kernel oops: http://paste.linux-vserver.org/256
1155852143 M * Hollow or should the persistant fix, fix that too?
1155852193 M * Bertl well, I was hoping that the persistent + helepr fixes this
1155852225 M * Hollow ok.. i didn't thought so after looking at the patches
1155852228 M * Hollow will test
1155852256 M * Hollow since i don't use persistent anyway, and the helper handling did not change, right?
1155852256 Q * ntrs Remote host closed the connection
1155852300 J * ntrs ntrs@68-188-51-87.dhcp.stls.mo.charter.com
1155852376 Q * meandtheshell Quit: bye bye ... 
1155852443 M * cehteh mhm .. doh .. no i am back .. and my kernel compile finished .. reboot soon again :P
1155852596 M * Bertl Hollow: i.e. use only the last two patches
1155852693 M * Bertl persist-fix01 + helper-feat03
1155852737 M * Hollow yeah, but i still don't see how it should affect the helper... beside the split of the persistent function method changed beside helper-feat02
1155852750 M * Hollow and i did not test it with persistent enabled
1155852755 M * Hollow so they even do not get called
1155852761 M * Hollow if i'd recompile
1155852773 M * Bertl I suspect you did hit the case I commented there
1155852783 M * Bertl i.e. the hashed/unhashed comment
1155852796 M * Bertl this was fixed with the restructuring there
1155852877 M * Bertl hmm, I'm probably wrong ehre, it is missing the shutdown
1155852879 M * Hollow ok, will take a sec to recompile + reboot
1155852902 M * Bertl I thought it would hit the hashed, but it did hit the shutdown
1155852953 Q * cehteh Ping timeout: 480 seconds
1155852954 J * ntrs_ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com
1155852959 M * Bertl Hollow: please add the following trivial change too
1155853017 M * michal_ Bertl: offtopic question - why you are not using some kind of version control system?
1155853026 M * Hollow heh.
1155853029 M * Bertl Hollow:  out_unhash:
1155853029 M * Bertl +       new_vxi->vx_state |= VXS_SHUTDOWN;
1155853029 M * Bertl         if ((vc_data.flagword & VXF_PERSISTENT))
1155853055 M * Bertl michal_: I do, I do :)
1155853062 M * Hollow ah, yes.. this makes sense
1155853078 M * michal_ i see. so it's just a private one ;)
1155853094 J * GNUcifer ~ct@cehteh.homeunix.org
1155853115 M * michal_ hey GNUcifer . cool nick :]
1155853136 M * Bertl welcome GNUcifer!
1155853150 M * Wonka hu cehteh
1155853165 M * GNUcifer heh .. my second nick since years ;)
1155853174 M * GNUcifer semms my wlan is fixed now
1155853393 J * lolilol hiddenserv@tor.noreply.org
1155853412 Q * ntrs Ping timeout: 480 seconds
1155853463 Q * olilo Ping timeout: 480 seconds
1155853508 M * Bertl GNUcifer: congrats!
1155853515 M * Bertl wb lolilol!
1155853548 M * GNUcifer mhm .. now moving out of sight of the AP ... maybe i loose connection again .. lets try
1155853570 Q * Hollow Quit: Konversation terminated!
1155853575 M * GNUcifer oh works
1155853588 M * GNUcifer never ever ipw2200 again
1155853625 M * GNUcifer the worst wlan chipset i'd ever used .. considering that the drivers are offically supported by intel it is utterly crap
1155853645 M * GNUcifer every hobbyist hacked driver for other chipsets has less problems
1155853665 M * Bertl interesting, work fine on morrigan's laptop
1155853674 M * doener works great for me as wekk
1155853679 M * doener s/wekk/well/
1155853701 J * comfrey ~comfrey@h-64-105-215-75.sttnwaho.covad.net
1155853707 M * Bertl wb comfrey!
1155853712 M * GNUcifer yes it works for many people .. but many others have problems
1155853725 J * Hollow ~hollow@2001:a60:f026::1
1155853727 M * GNUcifer you are just lucky when it works
1155853745 A * Bertl feels lucky ... :)
1155853751 M * michal_ Bertl: btw2
1155853766 M * michal_ my simple enter.c app is able to enter that suse vserver
1155853776 M * michal_ just vserver enter is not ;)
1155853790 M * GNUcifer and it changes with each new driver version .. be careful when updateing .. newer doesnt mean better with that
1155853804 M * Bertl daniel_hozac: any ideas regarding the suse guest/error?
1155853820 M * Bertl daniel_hozac: could it be that the tools get the startup wrong somehow?
1155853843 M * daniel_hozac Bertl: fairly unlikely if it's just using the sysv initstyle, it just runs rc.
1155853850 M * michal_ e-e
1155853855 M * michal_ pain initstyle here
1155853858 M * michal_ *plain
1155853864 M * michal_ pain too ;)
1155853869 M * Bertl ah, didn't I suggest to use sysv last time?
1155853878 M * michal_ yes, and it wasn't working at all
1155853885 M * Bertl a, hmm, means?
1155853917 M * michal_ that "unable to get vserver xid"
1155853924 M * michal_ scrool up, i cannot reboot now
1155853937 M * Bertl but even with plain, there is no reason for the tools to consider the guest not-running, when it is running, no?
1155853938 M * michal_ like rc would exit prematurly
1155853946 M * michal_ yes, you are right...
1155853962 M * daniel_hozac Bertl: depends, if the startup fails, it will clean up after itself.
1155853982 M * Bertl but that seems not to be the case, otherwise there would be no processes, right?
1155854017 M * michal_ startup suceeds actualy - half of it
1155854023 M * Bertl (or if I'm wrong, then the tools should kill off the remaining processes, when they consider the context gone, but it isn't)
1155854034 M * daniel_hozac how do i install a SuSE guest?
1155854036 M * michal_ i cannot see any of processes from 3 init lvl
1155854051 M * michal_ to be started. but i can enter that vserver anyway
1155854068 M * michal_ i won't tell you how...had just moved suse from main system into guest ;)
1155854102 M * Bertl okay, I guess for a test, an extremely reduced guest would suffice
1155854128 M * Bertl i.e. maybe try to remove everything but the libs and init itself (with config files)
1155854130 M * michal_ interesting...
1155854147 M * Bertl at least init must be running inside with plain, no?
1155854151 M * michal_ ssh from that host fails, because it cannot open /dev/tty
1155854159 M * michal_ from guest i mean
1155854164 M * michal_ yes, init is running
1155854185 M * michal_ dev/nulll is perm denied also
1155854195 M * daniel_hozac so your /dev is messed up.
1155854197 M * Bertl so try to make a copy of that guest (cp -va) and remove everything but init and init deps (libs)
1155854208 J * Vudu 7e70ad288f@perverz.hu
1155854212 Q * Vudu 
1155854227 M * Bertl daniel_hozac: should a messed up guest init keep you from entering?
1155854232 M * Hollow Bertl: semi-working.. if the configured vshelper does not exist it still returns with 0 (dmesg says: vxD: /foo: (startup 43) returned sync with 0)... if the vshelper does exist, and fails ENOEXEC is returned, but dmesg says: /usr/sbin/vshelper: (startup 43) returned sync with 256, while vshelper did exit(1)
1155854237 M * michal_ my dev is just that dev that was created with 'skeleton'
1155854271 M * Bertl Hollow: okay, the first case is IMHO okay, what do you think?
1155854286 M * daniel_hozac i think the first case is ok too.
1155854291 M * daniel_hozac we don't want to make vshelper mandatory...
1155854296 M * Bertl precisely
1155854298 M * coocoon michal_:  if u have a look into dev there are more files now aren't there
1155854306 M * Hollow well, with sc_helper we want it, no?
1155854330 M * daniel_hozac Hollow: probably the exit code is just << 8'd. there are macros to get it, IIRC.
1155854348 M * Bertl Hollow: yes, but trivial logging from the helper side should expose that
1155854357 M * Hollow ah, does call_usermodehelper return status like waitpid?
1155854379 M * Bertl yes, it also returns meta errors 
1155854408 M * Hollow ok, then the vxdprintk makes sense as-is
1155854408 M * Bertl e.g. non executable helpers should have different errors
1155854450 M * Hollow Bertl: how should the helper log if it does not exist? ;)
1155854468 M * Bertl Hollow: heh :) well, it _can_ log _if_ it exists :)
1155854484 M * Hollow but if it doesn't and sc_helper is enabled, we need to fail imo
1155854484 M * Bertl a missing log entry will easily expose a missing helper then
1155854494 M * daniel_hozac Hollow: you can have your starting process do some sanity checking, like util-vserver.
1155854526 M * Bertl let me check the helper details ...
1155854581 M * Hollow daniel_hozac: so every time methods who will (probably) result in a helper call should check if the helper is configured correctly?
1155854622 M * michal_ interesting. ircing now from suse guest, with kindly help of my enter app and bind mount of dev :P
1155854631 M * Bertl Hollow: well, thing is, a return of '0' means fine, no?
1155854638 M * michal_ so mayby it's that dev is messed up?
1155854664 M * Hollow Bertl: yep.. the question is, why does callusermodehelper return with 0 if kenrel-vshelper does not exist?
1155854676 M * Hollow *kernel.vshelper
1155854687 M * Bertl to what did you set it?
1155854693 M * coocoon michal_: i've said to u hours ago to have a look at dev's folder and look for the initscripts which messed it up
1155854693 M * Hollow  /foo
1155854717 M * michal_ coocoon: looks like suse is mounting 10 MB tmpfs over it!
1155854723 M * michal_ huh
1155854750 M * Bertl michal_: how can suse do that without the proper caps?
1155854753 M * coocoon michal_: also if u used the copy of ur system there will be a lot of more devs available as on my vserver created guests
1155854782 M * michal_ Bertl: uhm..indeed
1155854793 M * michal_ coocoon: nah
1155854800 M * michal_ only copied binaries and libs and etc
1155854804 M * michal_ and var
1155854820 M * coocoon initscripts are in etc
1155854828 M * Bertl you should use a skeleton's /dev for your guests
1155854847 M * coocoon i know that and i also must have a look for disabling a lot of scripts but i haven't done that before
1155854849 M * coocoon bertl
1155854851 M * Bertl Hollow: so I do not see a sane way to check for helper existance here
1155854864 M * michal_ i have used skeleton dev for this guest
1155854877 M * coocoon Bertl: if u use skeleton and start the suse guest there will be created new devs 
1155854888 M * coocoon in the devs folder
1155854889 M * Bertl I definitely doubt that
1155854891 M * coocoon thats suse
1155854892 M * Hollow Bertl: mhm, is callusermodehelper buggy, or is it intended?
1155854906 M * Bertl coocoon: a guest doesn't have the caps to create device nodes
1155854938 M * michal_ yes, it does not have
1155854948 M * coocoon bertl u can download a suse guest of mine have a look at ur dev folder and then start the guest and there will be more devs available 
1155854967 Q * DreamerC_ Quit: leaving
1155854984 J * DreamerC ~dreamerc@59-112-7-56.dynamic.hinet.net
1155854986 M * Bertl coocoon: I really doubt that, but I can try :)
1155854994 M * coocoon ok wait a moment
1155855023 M * Bertl Hollow: you can try to change the .retval = 0, to -1 in call_usermodehelper_keys()
1155855047 M * Hollow ok, will try
1155855172 M * derjohn Bertl, you should bet against cocoon  ;)  
1155855207 M * coocoon bertl: http://dev.qkon.eu/misc/suse_10.0_template.tar.bz2, i have make the tarball to 90 mb, i think more is not possible with yast ;-), maybe u must befor u start it have a look at the dev folder of this tarball because of i created it without cleaning it up
1155855549 M * Hollow Bertl: shouldn't execve in ___call_usermodehelper return ENOENT?
1155855735 Q * mire Quit: Leaving
1155855803 M * Hollow ah.. no it does not
1155855806 M * michal_ Bertl: just have removed dev from suse, copied a new one, from skeleton
1155855820 M * michal_ Bertl: and made ls before and after vserver suse start
1155855825 M * michal_ it differs lol
1155855860 M * Hollow it stores it in retval, but does exit(0) and sys_wait4 in wait_for_helper overwrites retval
1155855888 M * michal_ dev/null with proper permissions is beeing deleted and recreated with a wrong ones for example
1155855915 M * michal_ same goes for dev/tty
1155855933 M * michal_ dev/zero deleted
1155855948 M * michal_ a few symilnks created but it does not matter
1155855965 M * michal_ and a fifo, initctrl
1155855968 M * michal_ initctl
1155855994 M * coocoon so u have a lot of to do to disable them ;-)
1155856009 M * michal_ two questions
1155856013 M * michal_ 1. where?
1155856020 M * michal_ 2. how it is possible?
1155856024 M * Hollow initctl is ok
1155856026 M * coocoon and after an update with yast u have to do the same ;-)
1155856028 M * michal_ suse does not ave CAPMKDEV
1155856150 M * derjohn huh? michal_ coocoon which VS Versions are you using?
1155856163 M * Bertl michal_: show me the ls -la afterwards please (i.e. upload)
1155856267 M * coocoon derjohn: the latest one for sure
1155856289 M * derjohn coocoon, devel or stable experimental ? rc29 ?
1155856337 M * coocoon derjohn: i must say i have worked with the guests month ago haven't tried it out with the one which i am using at this moment
1155856362 M * coocoon so bertl have u tested it
1155856378 M * michal_ pastebin.de/11089
1155856380 M * derjohn but michal_ which VS version are you using ?
1155856399 M * michal_ derjohn: scrool up, been pasting it here today numerous times...
1155856459 M * Bertl no devices were created in that log
1155856474 M * Bertl null and tty did become 'normal' files
1155856490 M * Bertl (which is obviously the result from some over-eager scripts)
1155856503 M * derjohn and initctl| is not a device node
1155856513 M * Bertl it's a pipe
1155856548 M * derjohn michal_, is there something inside the dirs ?
1155856558 M * derjohn michal_, i mean: a device node?
1155856568 M * Bertl so while this probably explains your issues, it does not violate guest permissions
1155856579 M * Bertl I would try the following 'trick'
1155856581 M * michal_ net,shm,pts -> empty
1155856599 M * Bertl make the two 'critical' devices immutable
1155856617 M * Bertl the guest will fail to remove them, and everything should be fine
1155856617 M * michal_ tty,console?
1155856620 M * derjohn Bertl, it is allowed to take away the "w" permission from a device within a guest?
1155856650 M * Bertl yes, changing permissions of the directory entries is fine
1155856672 M * Bertl you can even rename/move them around 
1155856757 M * michal_ Bertl: chattr?
1155856759 M * michal_ +i
1155856769 M * michal_ asking, because have set a few
1155856781 M * Bertl yep
1155856793 M * michal_ and lsattr says that operation is not supported when reading flags on say,tty
1155856811 M * Bertl try with the setattr tool, or attack the dir itself :)
1155856830 M * michal_ setattr has iunlink
1155856839 M * michal_ butit will allow guest to unlink few devs :)
1155856847 Q * Hollow Quit: Konversation terminated!
1155856891 M * Bertl damn, well, I guess you need to hack the tools then, the kernel interface supports immutable too
1155856907 M * michal_ uh
1155856916 M * michal_ i'll try it with iunlink now and will see :)
1155856931 M * Bertl yes, but use iunlink-but-not-immutable
1155857033 M * coocoon so bertl if we would have bet i would lost it right
1155857055 J * Hollow ~hollow@2001:a60:f026::1
1155857077 M * Bertl coocoon: I'd say so :)
1155857093 M * coocoon so i will never bet with u about linux details ;-)
1155857107 M * coocoon or vserver details
1155857113 M * Bertl lol
1155857247 M * Bertl okay, I'm off for today ... will look into an improved helper solution tomorrow
1155857281 M * Bertl have a good one everyone, and cya tomorrow!
1155857286 N * Bertl Bertl_zZ
1155857348 M * michal_ ok
1155857363 M * michal_ does not work still, despite more and more changes
1155857373 M * michal_ i'll stay with my precious debian guests
1155857402 M * michal_ novell: learn how to make an os which is operating like it should be
1155857406 A * Belu is away (iŽll be back later...)
1155857407 N * Belu Belu_zZz
1155857412 M * michal_ and don't enclose your suse in its private namespace
1155857424 M * coocoon michal_: using grep for finding out which scripts do this
1155857443 M * michal_ coocoon: i have already deleted a lot of scripts
1155857450 M * michal_ and greped even more
1155857455 M * coocoon michal_: then u can post a workaround for me ;-)
1155857463 M * michal_ and i'm using only debian in production anyway
1155857479 M * michal_ coocoon: if i would have one ;)
1155857490 M * michal_ i give up on vservered-suse
1155857515 M * michal_ ok, cya, thx you all for help :)
1155857531 M * coocoon michal_: if u have a faster internet connection maybe try out my guest it is smaller