1104019319 M * Schak two friends are new to linux... and they want to "play" a bit... so i dont want to give them a shell to my dedicated root server... i had the choice between chroot jail and vserver... i decided for vserver ;)
1104019367 M * Schak i think they will install an ircd, psybnc... nothing more ;)
1104019386 M * Schak perhaps a LAMPP
1104019418 M * Bertl ah, excellent! 
1104019508 M * Schak are you the owner of 13thfloor?
1104019519 M * Bertl yep, I am ..
1104019539 M * Schak oh... master hisself *bow* :)
1104019544 M * Schak great work
1104019553 M * Bertl thanks!
1104019579 M * Bertl but it's a community project, so not all my work ...
1104019597 M * Bertl for example the tools are done by enrico (which isn't here atm)
1104019606 M * Schak how many ppl belongs to the core?
1104019611 M * Schak -s
1104019671 M * Bertl I'd say two hard core developer and a few others doing important things like testing or bithing at the devels or handling ml stuff ...
1104019706 M * Schak wow
1104019762 M * Bertl (you can find a complete list on the wiki)
1104019821 M * Bertl and feel free to help (for example by adding important stuff to the wiki or making suggestions how to improve linux-vserver)
1104020019 M * Schak yes, ok... if a should find a bug... or something i miss...
1104020064 M * Bertl also feel free to add yourself to the list of happy vserver users ;)
1104020573 M * Schak hehe... where do i find it?
1104020603 M * Bertl http://linux-vserver.org/VServer+Users
1104020636 M * Bertl (it's on the frontpage, under Public Relations)
1104022198 M * Schak i am searching for 20 mins... and i cant find it
1104022220 M * Bertl http://linux-vserver.org/
1104022252 M * Bertl scroll down, about in the middle of the page is:
1104022256 M * Bertl Public Relations
1104022258 M * Schak VServer Users Happy Linux-VServer Users 
1104022261 M * Schak ahhh %)
1104022263 M * Bertl and there is the link ... ;)
1104022274 M * Schak oh jesus... i should go to bed
1104022347 M * Doener Bertl: remember the mail you bounced to me about the chroot escape stuff on Nov 22?
1104022376 M * Bertl yep ..
1104022394 M * Doener back then you considered that one harmful (and IIRC I agreed)... we did the same now with the patch you did for the problem reported by jack
1104022419 M * Bertl right, that was why I was asking you ;)
1104022461 M * Doener so we agree that it actually wasn't harmful, right? because the necessary checks are done earlier in the ioctl handler...
1104022470 M * Bertl I hope so ...
1104022527 M * Doener ok. i trust cscope on that one ;)
1104022554 M * Bertl flags = flags & EXT2_FL_USER_MODIFIABLE
1104022561 M * Bertl this and similar should protect us
1104022567 M * Doener yes
1104022676 M * Schak i have a last small question before i go to bed:
1104022677 M * Schak # vserver base start
1104022677 M * Schak /sbin/vserver: line 543: ip: command not found
1104022677 M * Schak /sbin/vserver: line 545: ip: command not found
1104022677 M * Schak /sbin/vserver: line 543: ip: command not found
1104022677 M * Schak /sbin/vserver: line 545: ip: command not found
1104022695 M * Bertl hmm, iproute2 ?
1104022707 M * Schak no
1104022728 M * Bertl # rpm -qf `which ip`
1104022728 M * Bertl iproute2-2.2.4-13mdk
1104022778 M * Schak kara:/# rpm -qf `which ip`
1104022778 M * Schak rpmq: no arguments given for query
1104022778 M * Schak kara:/# which ip
1104022778 M * Schak kara:/# 
1104022803 M * Doener apt-get install iproute
1104023131 M * Schak outchy... my host system is unreachable... i think the "halt -f" in the vserver halted the host system
1104023132 M * Schak http://nopaste.chillfactory.net/show.php?id=138
1104023138 M * Schak here the last lines... 
1104023254 M * Bertl it's more likely that your vserver changed the networking stuff
1104023331 M * Bertl we had this when folks compiled debian kernels (which have the capability system as module and do not load them)
1104023361 M * Bertl basically a vserver can not reboot or halt the host ...
1104023377 M * Schak shouldnt work... because there is a dhcp client running... which gets the ip every reboot
1104023379 M * Schak hm ok
1104023424 M * Bertl the halt/reboot is redirected to the vshelper which acts in behalf of the vserver ...
1104023449 M * Schak i am waiting till my server got rebootet... then it should work... args... perhaps not, if the vserver stuff loads automatically at boot
1104023468 M * Doener only if you told it to do so
1104023493 M * Schak the "yes" is default in the debian-newvserver.sh :(
1104023505 M * Schak which creates the conf
1104023530 M * Doener but debian-newvserver.sh is for legacy tools, the configuration created by that one doesn't affect your vserver...
1104023543 M * Doener (this is why you had to create it by hand ;)
1104023551 M * Doener it = new style vserver config
1104023557 M * Schak ;)
1104023710 M * Schak and where is it in the new vserver config?... the start-on-reboot setting
1104023753 M * Doener flower page -> search for 'mark'
1104023909 M * Schak yes... found... hmm... dont know, if this file exist or not... ;)
1104023987 M * Schak strike... server is ready... and i am logged in via ssh 
1104024029 M * Bertl could you upload your kernel .config now please? ;)
1104024177 M * Schak oh yes... sorry. i thought, it wasnt interesting anymore... wait a min... i give you a link
1104024394 M * Bertl Doener: what do you think, should we force in the capability stuff?
1104024440 M * Doener well, we actually heavily depend on it, so i see no problems with that
1104024518 M * Schak http://213.23.239.158/schak-debian-2.6.9
1104024558 M * Doener hm, caps are built-in
1104024561 M * Bertl yep
1104024580 M * Doener hm, what is CONFIG_SECURITY_NETWORK ?
1104024584 M * Bertl Schak: why the decision to use INOXID_INTERN? (just curious)
1104024684 M * Bertl and after the 'base start' you lost connectivity?
1104024686 M * Schak whats that? without the "make menuconfig"-help i am helpless
1104024708 M * Schak yes... 2 mins ago the same...
1104024713 M * Bertl that's one of the xid tagging options ... UID/GID24 is the default
1104024772 M * Bertl okay, let me subsume that:
1104024780 M * Bertl you had the server working fine ...
1104024788 M * Bertl then you installed iproute2 (which was missing)
1104024803 M * Bertl then a new restart takes your system 'offline'
1104024815 M * Bertl is that correct so far?
1104024840 M * Schak i installed iproute... (apt-get install iproute) ... is iproute2 the same?
1104024872 M * Bertl yep
1104024875 M * Schak ok
1104024881 M * Schak and i didnt reboot
1104024885 M * Schak i startet the v
1104024896 M * Schak and 20 secs after restart
1104024898 M * Bertl yeah, I have your lines here ...
1104024902 M * Schak my host isnt reachable
1104024942 M * Bertl so I'd conclude that the vserver (maybe the config) is doing something with the ip (of your host) which confuses or disables routing ...
1104024958 M * Schak yes... i think so
1104024977 M * Bertl so as I further assume that you do not have any serial console to that host ...
1104025006 M * Bertl I'd suggest replacing the ip (from iproute2) with a script, and check the arguments
1104025024 M * Schak yes i have... a rescue system... and i can reboot via webinterface... no problem... in 10 mins my host is ready for the next test ;)
1104025183 P * Thorsten Leaving
1104025714 M * Doener Schak: do you use your 'main' ip address for the vserver?
1104025729 M * Doener if so, you have to tell the tools that the address is not to be touched
1104025894 M * Schak no... i have 4 different ips
1104025924 M * Bertl hmm ... would you care to show them to me in private?
1104025931 M * Bertl s/care/mind/
1104025931 M * Schak i have the command from the vserver base start.... concerning /bin/ip
1104026972 J * Thorsten ~Thorsten@dsl-084-057-024-066.arcor-ip.net
1104026995 M * Bertl wb Thorsten!
1104027009 M * Thorsten :-)
1104027164 M * Bertl Doener: hmm, have you spent a thought on allowing strace/ptrace from the host or xid=1 across chcontext()?
1104027182 M * Doener not yet...
1104027187 M * Bertl (do you think that is something worth some efford?)
1104027238 M * Doener may help with tools development, otherwise i can't come up with some real world use
1104029295 M * Bertl okay, interested in looking into the capability stuff in general?
1104029349 M * Doener which one? the masking stuff or making capabilities a requirement?
1104029368 M * Bertl if you ask me that way ... both of course ;)
1104029427 M * Doener hehe
1104029432 M * Bertl rs already verified that the basic idea seems to work ... but we still have to finish it (regarding the mask stuff)
1104029602 M * Bertl I'll update my todo list, feel free to add/change stuff there ...
1104029632 M * Bertl hmm, better I rename it to Todo List kernel ...
1104029634 M * Doener i'll look into that tomorrow, i'm on my way into bed now... but as i said i'll take at least a mid-deep look into everything ;)
1104029728 M * Doener you'll probably have to tell me about the masking stuff again, since 'our' 'discussion' faded away quickly last time, cause i got distracted :/
1104029749 M * Doener anyway, i'm tired right now, so let's get this going tomorrow
1104029753 M * Doener good night!
1104029782 M * Doener s/tomorrow/whenever you have time for that/ ;)
1104029799 N * Doener Doener_zZz
1104030485 M * Bertl okay, night!
1104030504 M * Thorsten night
1104034165 Q * Schak Read error: Connection reset by peer
1104034232 J * Schak schak@dsl-213-023-243-081.arcor-ip.net
1104034750 Q * Thorsten Ping timeout: 480 seconds
1104036677 Q * nox Ping timeout: 480 seconds
1104036678 J * nox ~vps@c150223.adsl.hansenet.de
1104036711 M * Bertl wb nox!
1104038858 M * Bertl okay folks, I'm off to bed now ...
1104038867 M * Bertl have a nice whatever everyone!
1104038882 N * Bertl Bertl_zZ
1104044402 Q * t Ping timeout: 480 seconds
1104044663 J * t ~tnichols@CPE-139-168-208-77.sa.bigpond.net.au
1104046144 Q * serving Read error: Connection reset by peer
1104047074 Q * click Ping timeout: 480 seconds
1104050556 J * click click@dsl-84-161.aal.tiscali.no
1104052812 J * JonB ~NoSuchUse@0x503e0321.kjnxx7.adsl.tele.dk
1104053106 J * serving ~serving@213.186.170.131
1104053632 Q * JonB Ping timeout: 480 seconds
1104056551 M * infowolfe argh, i hate dspam sometimes ;-)
1104056604 J * JonB ~NoSuchUse@0x503e0321.kjnxx7.adsl.tele.dk
1104056797 Q * _are_ Quit: Disconnecting
1104060023 J * t_ ~tnichols@CPE-139-168-209-129.sa.bigpond.net.au
1104060174 Q * flock Ping timeout: 480 seconds
1104060422 Q * t Ping timeout: 480 seconds
1104060428 J * q ~cumol@squid.dmz.fi
1104060433 M * q moin
1104060438 M * q anyone alive?
1104060481 M * infowolfe q, i am
1104060485 M * q fine
1104060491 M * q a question about vserver and ipv6
1104060495 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il
1104060504 M * infowolfe i'm not qualified to answer probably, but i'll give it a shot if i can
1104060528 M * q i've working ipv6 on the "root", but how to bind ipv6-ips to the vserver?
1104060574 M * JonB i dont think there is any IPv6 support
1104060609 M * infowolfe JonB, i'm thinking the same
1104060614 M * infowolfe Jonb, do you know much about dspam?
1104060620 M * JonB nothing
1104060636 M * q :(
1104060638 M * q suxx
1104060658 M * JonB q: i dont think anyone has had a real use for it yet
1104060665 M * JonB q: but... dont worry
1104060674 M * JonB q: you can most likely work arround it
1104060674 M * q JonB: i have a real use :)
1104060678 M * q JonB: how?
1104060701 M * JonB q: create aliases for your eth0
1104060714 M * JonB q: give them different IPv6 adresses
1104060734 M * JonB make other aliases for the vservers, and give those IPv4 address
1104060759 M * JonB then you use iptables to make a static NAT between the 2 adresses
1104060771 M * q uh. sounds complicated... there's no better way?
1104060787 M * JonB q: ask bertl, he's the main coder
1104060796 M * JonB q: or the mailing list, or the wiki
1104060804 M * q k
1104060806 M * q thnx. ill be back
1104060808 Q * q Quit: 2l8
1104060812 M * JonB q: and no, it's not that complicated
1104060844 Q * t_ Ping timeout: 480 seconds
1104063284 J * Thorsten ~Thorsten@dsl-084-057-024-198.arcor-ip.net
1104063734 J * t_ ~tnichols@CPE-139-168-209-155.sa.bigpond.net.au
1104064573 Q * Thorsten Remote host closed the connection
1104067737 J * Thorsten ~Thorsten@dsl-084-057-024-198.arcor-ip.net
1104068134 Q * sannes Read error: Connection reset by peer
1104068503 Q * DuckMaster Remote host closed the connection
1104069186 J * duckx ~Duck@dyn-83-157-202-15.ppp.tiscali.fr
1104069376 J * monrad ~monrad@213083190130.sonofon.dk
1104069479 Q * flock Ping timeout: 480 seconds
1104069580 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il
1104070554 J * dominance dominance@nyx.verfaction.de
1104070556 M * dominance hi
1104070574 M * dominance what's the use of /usr/include/vserver.h in the compiled package?
1104070577 M * dominance is it needed?
1104070581 M * dominance (of util-vserver)
1104070594 A * dominance is just trying to compile a util-vserver alpha debian package..
1104070729 Q * JonB Ping timeout: 480 seconds
1104070773 Q * duckx Remote host closed the connection
1104071214 J * duckx ~Duck@dyn-83-157-202-15.ppp.tiscali.fr
1104073201 Q * Thorsten Remote host closed the connection
1104075106 J * sannes ~ace@home.skarby.no
1104075349 Q * sannes Read error: Connection reset by peer
1104075476 Q * duckx Remote host closed the connection
1104075736 J * duckx ~Duck@dyn-83-157-202-15.ppp.tiscali.fr
1104075914 M * ndim dominance: Want my debian package?
1104076063 M * dominance ndim: have one already..
1104076079 M * dominance with consistent paths and against dietlibc as configure requested..
1104076099 M * dominance also cleaned up the debian/rules and added missing build-deps for the make docs..
1104076127 M * dominance basically i have now the spec instructions as far as i have been able to piece together which targets do what..
1104076547 J * JonB ~NoSuchUse@0x503e0321.kjnxx7.adsl.tele.dk
1104078065 M * ndim AFAICS, util-vserver still needs a lot of work to get from hack quality to production quality.
1104078094 M * JonB ndim: what is needed ?
1104078175 M * ndim Make it clear what is still needed, what is supposed to be in what path, what is needed on host systems, what on guest systems, what is legacy.
1104078179 M * ndim Stuff like that.
1104078193 M * ndim Oh, and a consistent configuration definition.
1104078197 M * dominance and a bit cleaned up configure/Makefile..
1104078207 M * JonB make it so
1104078207 M * dominance the hack to get the paths changed is ugly..
1104078233 M * dominance you need to define both lower and uppercase vars.. like so: "pkgstatedir=/var/run pkglibdir=/var/lib/util-vserver pkgstaterevdir=/var/run/vserver/.rev vshelperstatedir=/var/run/vshelper PKGSTATEDIR=/var/run PKGLIBDIR=/var/lib/util-vserver PKGSTATEREVDIR=/var/run/vserver/.rev VSHELPERSTATEDIR=/var/run/vshelper"
1104078255 M * ndim From what I've seen so far, e.g. the network config stuff needs serious changes to be extensible for IPv6 and other stuff.
1104078278 M * dominance ndim: even with NGN?
1104078291 M * JonB ndim: well i dont think anyone has asked for it yet
1104078303 M * ndim dominance: NGN?
1104078413 M * dominance next-gen networking
1104078419 M * dominance the thing you need for IPv6 in vserver
1104079030 N * Doener_zZz Doener
1104079037 M * Doener morning!
1104079067 M * dominance hi doener..
1104079910 J * infowolf1 ~infowolfe@mail.xhcl.net
1104080101 P * infowolf1 
1104080224 M * dominance doener: you use debian, don't you?
1104080499 M * Doener yes
1104080536 M * Doener currently there's gentoo on the box in front of me, but i've got access to some debina boxes
1104080542 M * Doener s/debina/debian/
1104080571 M * dominance hmmm
1104080595 M * dominance you're interested in helping ndim and me creating a good new debian deb by trying it?
1104080604 Q * t_ Ping timeout: 480 seconds
1104080627 J * t_ ~tnichols@CPE-139-168-209-155.sa.bigpond.net.au
1104080817 M * Doener did you contact the debian maintainer(s)? we're approaching a new try of cooperation between the debian folks and the vserver folks as you may have read on the ml
1104080830 M * dominance i read it..
1104080842 M * dominance and my idea is rather to propose a patch known to work for the community
1104080852 M * dominance than to have the DD slowly find out the hard way
1104080860 J * snapple ~blegha@217.238.33.65.cfl.rr.com
1104080868 M * dominance thus preparing a new deb and handing in a patch would make more sense IMHO
1104080892 M * dominance maybe you have followed some of the bugs in the Debian BTS.. then you should know how much engaged he's in getting the package straight..
1104080901 M * snapple if I use the grsec patches on the vserver host, will the vservers themselves be grsec'ed, since I really don't want that to happen
1104080907 M * dominance so the more you can come up with an "ANSWER" rather than a question, the besser..
1104080976 M * Doener didn't check the bug reports lately as i only use upstream stuff
1104081016 M * dominance well, there's a whole bunch of warnings and flaws in the alpha tools apparently..
1104081082 M * Doener but i guess if we can come up with a way that allows a) the debian folks to be happy (policy and stuff...) and b) the upstream folks to be happy (keep the packages as upstream as possible) cooperation should work out better than it currently does...
1104081143 M * dominance well, that's what the idea was so far =)
1104081146 M * snapple is it possible to install different distributions for vservers other than that of the host?
1104081156 M * Doener snapple: yes, it is
1104081157 M * dominance snapple: yes
1104081196 M * snapple I have seen examples using debbootstrap and it's redhat counterpart
1104081197 M * dominance Doener: so the point is we should at least sort the things out upstream that are in the way for a good packaging to happen
1104081205 M * snapple but what about if I want to install slackware
1104081213 M * snapple should I just run the install program from the cd?
1104081235 M * dominance Doener: and thus preparing a somewhat updated util-vserver source with a ready to deploy debian package will be the quickest way to convince Ola..
1104081281 M * Doener snapple: if that can create an installation in a chroot environment, yes... otherwise you'd have to figure out how to do that... easiest way may be to install it on another box and copy over the whole thing...
1104081325 M * snapple ok
1104081376 M * snapple if the host is grsec'd, will the vservers be grsec'd?
1104081385 M * snapple the host kernel
1104081552 M * Doener snapple: i guess so, vserver is not grsec aware and vice-versa
1104081638 N * Bertl_zZ Bertl
1104081650 M * Bertl morning folks!
1104081657 M * dominance hi Bertl ;)
1104081718 M * Bertl wow looks like there has been some discussion today ... without me ;)
1104081724 M * dominance hehe
1104081732 M * Doener dominance: just from my personal experience, i'd prefer to have Ola involved as early as possible, not mandatory as an active developer (although that would be nice), but at least discussions about the 'how and why' should be made between him and 'the others'. Both sides would be more aware of the needs of the others...
1104081736 M * dominance well, the "Debian issue" should be resolved.. once for all
1104081745 M * Bertl anybody willing to give a short conclusion?
1104081779 M * dominance Doener: you have mentioned "The Debian Policy" already.. if that one is respected, Ola will be happy..
1104081815 M * dominance Doener: the last thing Ola does want is more work than seriously neccessary.. and the more we can prepare the 0.30.197 to be "upstream convenient" that's not only helping Debian, but everyone out there..
1104081842 M * dominance Bertl: no conclusion.. util-vserver is not yet cleaned up and we need someone to make it "packageable"..
1104081867 M * dominance Bertl: like have the paths centrally configurable and to have the entire config files a bit more readible..
1104081873 M * JonB Bertl: ndim wants IPv6
1104081884 M * JonB Bertl: or maybe it was someone else
1104081890 M * JonB Bertl: but someone asked about it
1104081892 M * dominance Bertl: currently you need to set aswell lowercase and uppercase environment constants to get is respected while make install
1104081898 M * dominance JonB: it was ndim.. 
1104081906 M * Bertl yea, I guess Ola is willing to work on the debian packages and we will try to find sane solutions in the new year ...
1104081923 M * dominance Bertl: problem isn't within the deb.. problem is at least partly upstream
1104081946 M * Bertl I'm pretty sure, enrico is willing to adjust some things for debian too ...
1104081951 M * dominance Bertl: but i have tried to come up with a cleaned util-vserver alpha deb... for a discussion basis..
1104081959 M * Bertl (he already did for many other distros)
1104081965 M * dominance Bertl: this is nothin "Debian-specific"..
1104081995 M * Bertl okay, I have to get some dinner first ... back in 30min (then for real)
1104082000 M * JonB dominance: okay, thanks
1104082001 M * dominance Bertl: yet the debian-specific would be to compile the "doc" target already as a preparation of the release tarball..
1104082021 N * Bertl Bertl_oO
1104082031 M * dominance Bertl: the graphvis is in non-free and would thus break with the Debian "main" location..
1104082133 M * ndim graphviz :)
1104082144 M * dominance ok, viz..
1104082146 M * dominance ;)
1104082317 J * sannes ~ace@home.skarby.no
1104082559 Q * sannes Read error: Connection reset by peer
1104083098 N * Bertl_oO Bertl
1104083118 M * Bertl okay, back now for real ...
1104083144 M * Bertl ndim: you want ipv6 asap?
1104083249 M * Bertl dominance: do you have patches for 'improvements' or 'debianization' which doesn't hurt other distros (for the tools)?
1104083295 M * Bertl JonB: thanks for relaying the message ...
1104083315 M * JonB Bertl: you're velcome
1104083332 M * JonB Bertl: does it support IPv6 ?
1104083349 M * Bertl ngnet? not yet, but it's pretty easy to add, just needs somebody to test
1104083355 M * JonB Bertl: i told him about a work arround using iptavles
1104083379 M * Bertl the legacy networking code probably never will support ipv6
1104083548 M * JonB Bertl: okay, then ndim could test it for you
1104083781 M * Bertl yes, probably ... we'll see, he'll speak for himself ...
1104083900 M * JonB Bertl: ofc
1104084133 M * Doener Bertl: how far is 1.9.4 away?
1104084168 P * snapple Leaving
1104084283 M * ndim Bertl: No, I don't need IPv6 ASAP. I'd just hate to have /etc/vservers/foobar/{netthis,netthat} for IPv4 settings and /etc/vservers/foobar/IPv6 for IPv6. I'd much prefer /etc/vservers/foobar/net/{common,ipv4,ipv6,anothernetfamily}.
1104084309 Q * flock Ping timeout: 480 seconds
1104084320 M * ndim I don't remember the exact pathes (it was so badly documented the last time I looked at it that I didn't get it to actually run at all).
1104084381 M * ndim I just mentioned IPv6 as an example where the config format could use a clean redesign before the alpha code moves to beta status.
1104084412 M * Doener AFAIK there's no ipv6 support in the configuration...
1104084417 M * ndim Exactly.
1104084438 M * ndim And the IPv4 support in the config just assumes that IPv4 will be everything everyone would ever need.
1104084449 M * ndim But I gotta run...
1104084493 M * Bertl the configuration isn't able to handle ngnet yet
1104084500 M * Bertl and it isn't even required to do so ...
1104084501 M * Doener the net configuration ... too fast
1104084535 M * Bertl but yes, we will need config support for ipv6
1104084632 M * dominance Bertl: re.. well i have started to try packaging for now..
1104084644 M * dominance Bertl: and the "improvements" aren't yet patchified..
1104084656 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il
1104084659 M * dominance Bertl: for i'd love to have someone "in the know" betatest the packs first..
1104084666 M * Bertl well, I'd suggest sending them to the ml (with cc to enrico) asap
1104084676 M * Bertl (in patch form)
1104084695 M * dominance Bertl: and once we have proven that the effect is assured, there'll be patches..
1104084709 M * dominance Bertl: well.. let's start with the compile errors..
1104084711 M * Bertl and if possible, separate debian specific patches from general improvements
1104084714 M * dominance ehrm.. warnings..
1104084720 M * dominance while building the util-vserver i get those:
1104084729 M * dominance http://backend.verfaction.de/~kk/util-vserver/buildlog_stderr.log
1104084755 M * dominance and have to define "pkgstatedir=/var/run pkglibdir=/var/lib/util-vserver pkgstaterevdir=/var/run/vserver/.rev vshelperstatedir=/var/run/vshelper PKGSTATEDIR=/var/run PKGLIBDIR=/var/lib/util-vserver PKGSTATEREVDIR=/var/run/vserver/.rev VSHELPERSTATEDIR=/var/run/vshelper" as environments
1104084787 M * dominance or maybe i didn't find the correct switches for configure
1104084790 M * Bertl the #warnings are developer output ...
1104084811 M * Bertl and they are _not_ gcc extensions, they are preprocessor extensions
1104084814 M * dominance the really annoying about these are that that i have to define them twice as lower- and as uppercase
1104084828 M * dominance "src/chbind.c:200: warning: warning: gethostbyname() leaks memory.  Use gethostbyname_r instead!" too? *g*
1104084849 M * Bertl no, but why is gethostbyname leaking memory?
1104084890 M * dominance i didn't put that warning there.. i just copied stderr =)
1104084922 M * Bertl yes, I know, ... thing is this is probably done with a 3.4 or 4.0 gcc
1104084925 M * dominance and from a user/packager point of view i don't care *WHY* they happen.. i just read their text.. =)
1104084932 M * dominance 3.3
1104084947 M * dominance gcc (GCC) 3.3.5 (Debian 1:3.3.5-5)
1104084949 M * Bertl interesting that complains about this too? sec
1104084950 M * dominance to be precise..
1104084976 M * dominance there is a gcc-3.4 on that host, but it's not the default gcc
1104084986 M * dominance thus if this one is chosen, then that's just another flaw of util-vserver
1104085038 M * Bertl I would opt against a 3.4 gcc for now ... it's too buggy ...
1104085042 M * dominance *ggg*
1104085049 M * dominance for amd64 it's recommended though
1104085067 M * Bertl well, it doesn't even compile the kernel ... *G*
1104085069 M * dominance so the gcc-3.4 should be allowed to be used at least..
1104085084 M * dominance mmh? for amd64 it does afair
1104085097 M * Bertl another question, how many warnings do you get when you compile the 'debian' kernel?
1104085119 M * JonB Bertl: with or without --pedantic and --ansi ?
1104085121 M * dominance well, let's put this gcc-3.4 aside.. for now it's not important.. it shouldn't be used in the first place..
1104085148 M * Bertl with the same option dominance used here but with a debian default .config
1104085172 M * dominance the vserver-info does output: 
1104085173 M * dominance              util-vserver: 0.30.196; Dec 26 2004, 15:36:22
1104085173 M * dominance Features:
1104085174 M * dominance                        CC: i386-linux-gcc, i386-linux-gcc (GCC) 3.3.5 (Debian 1:3.3.5-5)
1104085174 M * dominance                       CXX: i386-linux-g++, i386-linux-g++ (GCC) 3.3.5 (Debian 1:3.3.5-5)
1104085174 M * dominance                  CPPFLAGS: ''
1104085174 M * dominance                    CFLAGS: '-Wall -g -O2 -std=c99 -Wall -pedantic -W'
1104085176 M * dominance                  CXXFLAGS: '-g -O2 -ansi -Wall -pedantic -W -fmessage-length=0'
1104085178 M * dominance                build/host: i386-pc-linux-gnu/i386-pc-linux-gnu
1104085180 M * dominance              Use dietlibc: yes
1104085188 M * dominance        Build C++ programs: yes
1104085190 M * dominance        Build C99 programs: yes
1104085192 M * dominance            Available APIs: compat,v11,v13,fscompat,net,oldproc,olduts
1104085193 M * Bertl and don't get me wrong, most of those 'warnings' can and will probably be fixed ...
1104085196 M * dominance             ext2fs Source: e2fsprogs
1104085198 M * dominance     syscall(2) invocation: fast
1104085201 M * dominance       vserver(2) syscall#: 273/glibc
1104085204 M * dominance oh, i'm sure they will =)
1104085215 M * dominance they don't make my primary problem anyway =)
1104085226 M * Bertl thought so ...
1104085256 M * dominance ndim had some thoughts too about splitting up the host and guest tools which make sense to me and some patches about manpages in the wrong category and stuff
1104085279 M * dominance so if someone using a Debian host and knowing vserver by heart could help debugging the debs, that'd help allot.
1104085307 M * dominance so ndim and me can do debs and when they are working like they should, then we could propose them upstream to Ola..
1104085308 M * Bertl what are the 'guest' tools?
1104085344 M * dominance uhm, haven't checked that indepth.. lemme check ndim's rules..
1104085401 M * dominance he has those in the guest tools package:
1104085402 M * dominance man/chbind.8
1104085402 M * dominance man/chcontext.8
1104085402 M * dominance man/reducecap.8
1104085402 M * dominance man/vps.8
1104085404 M * dominance man/vpstree.8
1104085405 M * dominance man/vserver-stat.8
1104085407 M * dominance man/vserver.8
1104085409 M * dominance man/vtop.8
1104085444 M * Bertl aha, and what are the host tools then?
1104085454 M * Doener hm, without chcontext you can as well leave the kernel unpatched :)
1104085459 M * dominance hmmm
1104085467 M * dominance interestingly the same..
1104085481 M * Bertl aha ;)
1104085481 M * dominance i guess i should build his package and check.. hold on a sec..
1104085712 M * dominance hmm
1104085747 M * dominance i guess ndim should explain that by himself..
1104085757 M * dominance but to me this looks like a "work in progress"
1104085815 Q * t_ Ping timeout: 480 seconds
1104085833 J * t_ ~tnichols@CPE-139-168-209-155.sa.bigpond.net.au
1104085842 M * dominance somewhat i can't really see the difference in the built tools.. just in the depends..
1104085873 M * Bertl okay, we'll wait for him ...
1104085889 M * Doener Bertl: any ETA for 1.9.4? trying to decide whether i should ask for a split patch for the current version or for 1.9.4 ;)
1104085906 M * Bertl well, you'll get a split very soon ...
1104085912 M * dominance interestingly though i can't change /etc/vservers/util-vserver-vars and have vserver-info tell me the new location of the vserver-root
1104085923 M * Doener sounds like 1.9.4 is near
1104085990 M * dominance hmm
1104086004 M * Doener dominance: did you change the path to util-vserver-vars? here it is below ..../lib/util-vserver 
1104086025 M * dominance DEFAULT_VSERVERDIR='/var/lib/vservers' is in /var/lib/util-vserver/util-vserver-vars, yet /etc/vservers/util-vserver-vars has VROOTDIR='/srv/vservers'
1104086049 M * dominance i wonder if this should be symlinked to /var/lib/util-vserver/util-vserver-vars
1104086058 M * Doener there's no /etc/vservers/util-vserver-vars at all on my box
1104086092 M * Bertl imho 90% of the issues are realted to bad configure time options
1104086097 M * dominance well, i don't install this especially.. so the make install or make install-distribution needs to be doing this
1104086118 M * dominance Bertl: that's why i'm asking that someone with vserver experience would shed some light onto this
1104086148 M * dominance Bertl: the configure is run with --with-vrootdir=/var/lib/vservers .. so it's correct to assume that for the start..
1104086158 M * dominance but once i change this, it should someone head to the new location
1104086161 M * Bertl util-vserver-0.30.196]# ./configure --help
1104086198 M * Bertl setting the various prefixes and such is probably sufficient ...
1104086216 M * Bertl it even displays your system defaults
1104086235 M * dominance Bertl: hmmm,  haven't seen options for what i wanted to change..
1104086255 M * Bertl   --with-initrddir <DIR>  use <DIR> as directory for SysV init-files (default:
1104086255 M * Bertl                           $sysconfdir/rc.d)
1104086255 M * Bertl   --with-kerneldir=DIR    assume kernelsources in DIR (default:
1104086255 M * Bertl                           /lib/modules/<current>/build)
1104086255 M * Bertl   --with-vrootdir=DIR     place vservers under DIR (default: /vservers)
1104086269 M * dominance sure.. nothin wrong with that..
1104086279 M * Bertl plus the config, bin etc dirs
1104086282 M * dominance but how to set: pkgstatedir=/var/run pkglibdir=/var/lib/util-vserver pkgstaterevdir=/var/run/vserver/.rev vshelperstatedir=/var/run/vshelper
1104086300 M * dominance pkgstatedir would else become /usr/var/run
1104086309 M * dominance pkglibdir would be /usr/lib/util-vserver
1104086312 M * Bertl   --localstatedir=DIR    modifiable single-machine data [PREFIX/var]
1104086327 M * Bertl   --libdir=DIR           object code libraries [EPREFIX/lib]
1104086329 M * dominance hmm, lemme check this..
1104086345 M * Bertl don't tell me that is the first time you use configure? ;)
1104086350 M * dominance nope =)
1104086370 M * dominance but a quick search thru Makefile.am and configure didn't yield my searchstrings..
1104086732 M * dominance btw. how tight is the need to have iproute and vlan in the depends?
1104086756 M * Bertl iproute is essential, vlan probably not if you don't have vlans
1104086783 M * dominance well, i still see a /usr/lib/util-vserver/FEATURES.txt with the localstate set
1104086819 M * dominance this the pkglibdir isn't set yet..
1104086838 M * dominance the rest is ok so far
1104086922 M * Bertl what did you set libdir too?
1104086931 M * Bertl --libdir=
1104086946 M * dominance well, i want the *LIB* part in /usr/lib.. but the file archive library in /var/lib/..
1104086956 M * dominance thus setting the libdir will not yield what i want..
1104086974 M * dominance or did i get something wrong?!
1104086997 M * Bertl well, you probably need a new option for that ...
1104087009 Q * flock Ping timeout: 480 seconds
1104087022 M * dominance ./usr/lib/libvserver.so.0.0.0 is ok in /usr/lib.. but the pkglibdir isn't supposed to be /usr/lib IMHO
1104087033 M * dominance thus pkglibdir is already sufficient for me..
1104087069 M * Bertl if that works ... not sure the separation is there ...
1104087100 M * dominance that'd be the next question =))
1104087120 M * dominance but at least i've tightened the build-deps on dietlibc already..
1104087136 M * dominance or ain't it valid anymore that dietlibc is *strongly* recommended?
1104087150 M * Bertl it is, because of the resolver issues
1104087167 M * dominance ok
1104087168 M * Bertl (you have to ask enrico for details)
1104087177 Q * JonB Quit: Leaving
1104087186 M * dominance (i don't want details.. i want answers.. solutions.. not more problems)
1104087192 M * Bertl is the dietlibc for debian already fixed?
1104087197 M * dominance fixed?
1104087200 M * Bertl (I mean for amd64)
1104087201 M * dominance was it broken?
1104087207 M * dominance amd64 is no official port
1104087213 M * dominance thus it can be broken and i don't care
1104087215 M * Bertl yes it was and maybe still is broken on 64bit archs
1104087230 M * dominance if on all 64bit archs, then that'd be a problem
1104087281 M * Bertl dominance: try to look a little beyond your 'usual' field of application ;)
1104087299 M * dominance :-P once i have things working which i need
1104087314 M * Bertl in the next year, linux-vserver will be used on amd64 more often than on x86 ...
1104087316 M * dominance but *first* i want a known-working.. then go for a generic-known-working..
1104087327 M * dominance and as my server is amd64 i will want amd64 working too, don't worry.. =)
1104087391 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il
1104087441 M * Bertl wb flock!
1104087673 M * dominance Bertl: dietlibc build problems on 64bit have been marked as resolved btw.
1104087693 M * Bertl excellent!
1104088023 M * Bertl Doener: http://vserver.13thfloor.at/Experimental/split-vs1.9.3.14/
1104088145 M * Bertl and http://linux-vserver.org/ToDo+List+Kernel ;)
1104088188 M * dominance .14? what's new?
1104088213 M * Doener thanks!
1104088213 M * Bertl a few cleanups, nothing essential ...
1104088242 J * _are_ ~are@dsl-082-082-062-022.arcor-ip.net
1104088250 M * dominance ok
1104088256 M * Bertl welcome _are_!
1104088299 M * Doener ah, i always wanted to ask what "reduce usage of current->*" means... replace those lines by using the existing macros?
1104088336 M * Bertl hmm, no not really, it's jsut the first step
1104088349 M * Bertl the thing is, current->* is expensive in the kernel
1104088359 Q * flock Ping timeout: 480 seconds
1104088368 M * Bertl so you want to reduce the number of invocations ...
1104088378 M * Doener i see
1104088404 M * Bertl for example doing xid_t xid = current->xid; and then using that is cheaper as calling current->xid twice
1104088417 M * Bertl (and not using it at all is the cheapest of course ,)
1104088533 M * Doener hmm... i still have to test my tagxid-remount patch...
1104088582 M * _are_ hi
1104088583 M * Bertl a few things have changed around the xid handling ..
1104088638 Q * Hollow Remote host closed the connection
1104089530 J * sannes ~ace@home.skarby.no
1104089565 M * Bertl welcome sannes!
1104089875 Q * t_ Ping timeout: 480 seconds
1104090262 J * t_ ~tnichols@CPE-139-168-209-155.sa.bigpond.net.au
1104090287 M * Bertl wb t_!
1104092367 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il
1104092422 M * Bertl wb flock!
1104092439 M * flock 'lo
1104093331 M * Bertl it's hard to believe but I finished a allmodconfig build for 2.6.10 ... (with vserver)
1104093342 M * Bertl du says:  581248 /src/modules/2.6.10-vs1.9.3.14
1104093393 M * Bertl (well okay it's with debug info included)
1104093550 Q * mcp Ping timeout: 480 seconds
1104093979 M * Schak hi, can anyone help me? -> i build 2.6.9 with uid/gid32 and build a vserver, which was working perfectly... few hours ago i switched to uid/gid24, rebuild the kernel, rebooted... and now exim has lots of errors and doenst start at all... strace tells me, that he cant access to /etc/passwd ... but this file has 644!
1104094018 M * Bertl hmm, did you mount any partition with tagxid?
1104094095 M * Schak no... there is nothing i mounted... the vserver is unter /vservers... and this is a path in /dev/hda3 (root)
1104094108 M * Schak ext3
1104094144 M * Bertl okay, the fstab for that mount doesn't contain tagxid or xid in the line for /vservers ?
1104094223 M * Bertl anyway, let's check those files with lsxid ...
1104094228 M * Schak yes, i havent touched the /etc/fstab at the host yet
1104094347 M * Schak cd /vservers/base/etc ... lsxid ... i get EVERYWHERE !!ERR!!
1104094363 M * Schak vserver base is running at the moment
1104094403 M * Bertl okay, that looks fine so far ...
1104094413 M * Bertl did you compile the linux-vserver debug stuff in?
1104094418 M * Schak no
1104094442 M * Schak shall i?... no problem, if it would help
1104094443 M * Bertl hmm .. okay, that would help here ...
1104094461 M * Schak ok... give me 30 mins
1104094472 M * Bertl btw, you should update to 2.6.10-vs1.9.3.14 if possible
1104094496 M * Bertl (it has some fixes in kernel and vserver)
1104094504 M * Schak oh ok... and really stable?
1104094527 M * Bertl looks good so far ... but the release candidates where pretty stable (kernel is two days old)
1104094626 M * Schak and... do you know, if the 2.6.10 will accept my 2.6.9 config completely?
1104094648 M * Schak i will try... ;)
1104094768 M * Schak [ ]   Limit the IDLE task
1104094770 M * Schak do i need this?
1104094795 M * Schak or: is it recommended to enable this?
1104094831 M * Bertl no, the default is fine ...
1104094858 M * Schak ok
1104094861 M * Bertl you just need that if you want precise hard scheduling even if the machine is mostly idle
1104094880 M * Bertl (which has a small overhead and is usually not required)
1104095673 M * Schak do u have the complete url, where i can get the vs1.9.3.14? or is it on the www.13thfloor.de ?
1104095696 M * Schak at... sry
1104095719 M * Bertl yep, sec
1104095744 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.10-vs1.9.3.14.diff
1104095845 M * Schak thx
1104097300 M * Schak this error came approx 50 times... how would you estimate the problem?
1104097300 M * Schak drivers/net/via-rhine.c:1636: warning: passing arg 1 of `readw' makes pointer from integer without a cast
1104097300 M * Schak drivers/net/via-rhine.c:1637: warning: passing arg 1 of `readw' makes pointer from integer without a cast
1104097300 M * Schak drivers/net/via-rhine.c:1647: warning: passing arg 2 of `writeb' makes pointer from integer without a cast
1104097300 M * Schak drivers/net/via-rhine.c:1662: warning: passing arg 2 of `writeb' makes pointer from integer without a cast
1104097924 M * Loki|muh hey guys
1104097962 M * Loki|muh do you think that CAP_FOWNER is not bad to give to a vserver?
1104097968 M * Loki|muh this is needed for acls
1104098008 M * Bertl Schak: that is fine .. the error is new, the problem is very old ...
1104098024 M * Bertl hey Loki|muh!
1104098084 M * Bertl Loki|muh: hmm, isn't that part of the default capabilities?
1104098147 M * Loki|muh I dont know... a customer came to me and asks for the capability
1104098175 M * Bertl sec, checking ...
1104098203 M * Bertl CapPrm:00000000d44c04ff
1104098203 M * Bertl CapEff:00000000d44c04ff
1104098216 M * Bertl so yes, that capability is given for vserver root by default
1104098256 M * Loki|muh ah thx
1104098259 M * Bertl the entire CAP_FS_MASK by the way
1104098272 M * Loki|muh in which file can I find this?
1104098278 M * Bertl include/linux/capability.h
1104098344 M * Bertl Schak: the various driver writers have to 'adapt' their drivers to use proper pointers ...
1104098380 M * Schak and this has not been done yet ... ok ;)
1104098382 M * Schak Linux kara 2.6.10-vs1.9.3.14 #1 Sun Dec 26 22:38:53 CET 2004 i686 GNU/Linux
1104098383 M * Schak works!
1104098401 M * Bertl what does kara refer to, if I may ask?
1104098402 M * Loki|muh Bertl: there is no CapPrm and CapEff in /usr/include/linux/capability.h
1104098427 M * Bertl that one is the output of chcontext --ctx 100 --secure grep Cap /proc/self/status
1104098427 M * Schak sorry? ... kara is my host
1104098442 M * Bertl yeah, does that name refer to anything?
1104098469 M * Schak yes... java programming... eth zürich... kara, der marienkäfer ;)
1104098486 M * Bertl ah, okay, i.c. thanks ;)
1104098491 M * Schak you know this?
1104098530 M * Loki|muh Bertl: ah thx! :)
1104098537 M * Bertl Schak: yes, I know it .. 
1104098558 M * Bertl EASA 2002 IIRC
1104098616 M * Schak context: "endliche automaten"... great stuff ;)
1104098625 M * Schak .o0( i think )
1104098641 M * dominance Bertl: when will ensc most probably show up next?
1104098642 M * Bertl yeah, I made my diploma thesis in the area of graph grammars ... ;)
1104098670 M * Bertl dominance: best you send him an email (better mail it on the ml, and cc him)
1104098694 M * dominance mmh, ok.. 
1104098701 M * Bertl (he is currently busy with his thesis)
1104099180 M * Schak welllll... my base vserver is running... debug mode is compiled into the kernel... what do i have to do now?
1104099211 M * Bertl try to start the vserver with exim (or exim)
1104099219 M * Bertl watch the messages in the syslog
1104099245 M * Schak syslog @ the V or @ the host?
1104099245 M * Bertl on what filesystem is taht exim?
1104099258 M * Bertl the syslog and kernel log on the host
1104099354 M * Schak concerning the vserver in kern.log only:
1104099356 M * Schak Dec 26 23:12:19 kara kernel: process `named' is using obsolete setsockopt SO_BSDCOMPAT
1104099360 M * Schak in syslog nothing
1104099393 M * Bertl but exim fails?
1104099394 M * Schak exim was tried to get started via init.d scripts when i typed vserver base start
1104099396 M * Schak yes
1104099412 M * Bertl okay, then it is 99% exim only related ...
1104099425 M * Bertl let's strace that bugger ...
1104099428 M * Schak interested in the strace?
1104099429 M * Schak ok
1104099606 M * Schak but i cannot explain, why exim was working perfecty... since i rebooted the host, where i changed the kernel to uid/gid24, this problem exist
1104099637 M * Bertl well, if you want to do a 'quick' verify on that, just change the tagging and try again ...
1104099652 M * Bertl (I'm pretty confident it is not related)
1104099684 M * Bertl nevertheless, it could be a vserver related issue, so let's investigate it a little further ...
1104099712 M * Schak http://nopaste.chillfactory.net/show.php?id=139
1104099743 M * Schak ok... if you cannot recognise any problems in the strace pasting, then i will change back to uid/gid32
1104099752 M * Bertl it's trying to connect to /dev/log ...
1104099755 M * ndim re
1104099795 M * Bertl Schak: check what /dev/log inside the vserver is, and if it is connected to the syslogd and if the permissions allow for a connect .. maybe you are using the wrong pipe?
1104099804 M * Bertl wb ndim!
1104099815 M * Schak /dev/log doesnt exist
1104099821 M * Bertl dominance: ndim is back ...
1104099849 M * Bertl Schak: maybe you didn't start the syslog service?
1104099853 M * Bertl (on the vserver)
1104099854 M * dominance apparently
1104099915 M * Schak aaaah... i got it... yesterday i removed the sysklogd... and the syslogd is included in the sysklogd package
1104099920 A * Bertl just thought you had some questions ... 
1104099952 M * ndim I'm going to start reading backlog now.
1104099952 M * Schak you were right... i told it... yes yes... 
1104099986 M * Bertl you can probably work around that (if you don't want syslogd inside the vserver) by replacing /dev/log with a /dev/null
1104100141 M * Schak /dev/log exist now... syslogd is started... but exim doesnt start... same error msg
1104100177 M * Schak base:/dev# /etc/init.d/exim start
1104100177 M * Schak Starting MTA: 2004-12-26 22:28:35 Exim configuration error
1104100177 M * Schak   group mail, referred to in line 213, was not found
1104100177 M * Schak 2004-12-26 22:28:35 Exim configuration error
1104100177 M * Schak   group mail, referred to in line 213, was not found
1104100178 M * Schak exim: could not open panic log - aborting: original error above
1104100239 M * Bertl well, does group mail exist in /etc/group ?
1104100260 M * Doener does /var/log/exim exist as a directory?
1104100279 M * Doener 258: open("/var/log/exim/paniclog", O_WRONLY|O_APPEND|O_CREAT, 0640) = -1 EACCES (Permission denied)
1104100304 M * Doener hm, and /etc/group isn't accessable as well...
1104100325 M * Schak base:/var/log/exim# ls -l
1104100325 M * Schak total 4
1104100325 M * Schak -rw-r-----    1 mail     adm           897 Dec 26 05:53 mainlog
1104100325 M * Schak -rw-r-----    1 mail     mail            0 Dec 26 05:25 paniclog
1104100341 M * Bertl is exim started as user mail?
1104100343 M * Schak base:/var/log/exim# ls -l /etc/group /etc/passwd
1104100343 M * Schak -rw-r--r--    1 root     root          441 Dec 26 04:32 /etc/group
1104100343 M * Schak -rw-r--r--    1 root     root          917 Dec 26 13:44 /etc/passwd
1104100364 M * Bertl ah, maybe exim is chrooting into some other place?
1104100387 M * Schak yes... exim is started with root rights
1104100398 M * Schak and i guess it ends as user mail
1104100431 M * Bertl could you update the strace (and do it with -fF) please?
1104100472 M * Doener setregid32(0xffffffff, 0x8)             = 0
1104100472 M * Doener setresuid32(0xffffffff, 0x8, 0xffffffff) = 0
1104100525 M * Doener user looks fine...
1104100534 M * Bertl I'm more interested in what /var/spool/exim is/should contain ...
1104100571 M * ndim My motivation to separate into -host and -guest was that there is stuff in there which only makes sense on a host system (stuff like starting a vserver), and stuff which only makes sense on a guest system (stuff like starting a daemon with a special binding). And you won't need one on the other.
1104100606 M * Bertl ndim: okay, but util-vserver doesn't contain _any_ guest tools, right?
1104100607 M * Doener ndim: you can't control bindings from within the vservers...
1104100614 M * Schak http://nopaste.chillfactory.net/show.php?id=140
1104100659 M * ndim Bertl: The version I downloaded some time ago did, as far as I could discern. If it doesn't, either I am quite insane or it is quite badly documented.
1104100662 M * ndim :)
1104100686 M * ndim Doener: I don't remember all the details. I did that stuff a few months ago.
1104100687 M * Bertl Schak: okay what does ls -la /dev/log inside the vserver give?
1104100717 M * Bertl ndim: hmm, okay, what 'tools' are considered guest stuff by you?
1104100737 M * ndim Lemme check my package...
1104100739 M * Schak base:/var/spool/exim# ls -la /dev/log
1104100739 M * Schak srw-rw-rw-    1 root     root            0 Dec 26 23:28 /dev/log
1104100764 M * Bertl and ls -lda /dev
1104100765 M * Schak http://nopaste.chillfactory.net/show.php?id=141
1104100780 M * Schak base:/var/spool/exim# ls -lda /dev
1104100780 M * Schak drwxr-xr-x    3 root     root        24576 Dec 26 23:28 /dev
1104100813 M * Schak ... missing w for all?
1104100830 M * Bertl no no that's fine ...
1104100833 M * Schak ok
1104100847 M * Bertl do you get any messages on the host (syslog or dmesg)?
1104100928 M * Schak dmesg at the V is not permitted... dmesg at the host gets nothing, what could help in this case
1104100959 M * Bertl okay, for your info, we are looking for vsW messages 
1104100985 M * Bertl btw, did I miss your answer regarding the filesystem your vserver is located on?
1104101117 M * Schak it is a simple path in the main root system, which is ext3... no "vserver in a box"
1104101171 M * Bertl okay, so we have to find out why connect via /dev/log doesn't work ... any ideas?
1104101222 M * Bertl Schak: try with logger
1104101270 M * Schak how does logger work?
1104101280 M * Schak which command do i have to type?
1104101287 M * Schak logger /etc/init.d/exim ?
1104101301 M * Schak logger /usr/sbin/exim ... ?
1104101353 M * Bertl logger -p local0.notice "Hello folks!"
1104101428 M * Bertl or logger -p kern.crit "Hello folks!"  if your syslog ignores the other ;)
1104101432 M * ndim Bertl: I found e.g. /usr/lib/util-vserver/fakerunlevel and /etc/init.d/v_* to belong to -guest.
1104101456 M * Bertl ndim: no, they are exclusively host specific (at least the v_*)
1104101460 M * ndim Bertl: If this is wrong, I'd like to ask you whether you are really using *nothing* vserver related in guest systems?
1104101475 M * Schak Dec 26 23:47:34 base logger: /etc/init.d/exim
1104101475 M * Schak Dec 26 23:50:06 base logger: Hello folks
1104101477 M * Schak hrhr ;)
1104101493 M * Bertl ndim: AFAIK we do not change anything inside the guest ...
1104101521 M * Bertl Schak: okay, now try with logger -p local0.notice -u /dev/log "Hello folks!"
1104101556 M * Bertl or maybe logger -p local0.notice -f /dev/log "Hello folks!" ?
1104101576 M * Schak base:/var/log# logger -p local0.notice -u /dev/log "Hello folks"
1104101576 M * Schak connect: Protocol wrong type for socket.
1104101613 M * Schak base:/var/log# logger -p local0.notice -f /dev/log "Hello folks"
1104101613 M * Schak logger: /dev/log: No such device or address.
1104101642 M * Bertl I've got another one:  logger -p local0.notice -d -u /dev/log "Hello folks!"
1104101720 M * Schak Dec 26 23:54:15 base <someone>: Hello folks
1104101724 M * Schak in syslogd
1104101733 M * Schak ehh... /var/log/syslog
1104101755 M * Bertl okay ... looks good, let's try that from another user ... do an su to the mail user ...
1104101784 M * Schak doesnt work... because he has shell /bin/false... i edit this... wait
1104101839 M * Schak hmm strange
1104101847 M * Schak base:/var/log# su mail
1104101847 M * Schak No shell
1104101856 M * Schak but: mail:x:8:8:mail:/var/mail:/bin/sh
1104101862 M * Schak in /etc/passwd
1104101898 M * Bertl does /var/mail exist?
1104101935 M * Schak base:/var/mail# ls -Ral
1104101935 M * Schak .:
1104101935 M * Schak total 12
1104101935 M * Schak drwxrwsr-x    2 root     mail         4096 Dec 26 05:48 .
1104101935 M * Schak drwxr-xr-x   15 root     root         4096 Dec 26 04:14 ..
1104101936 M * Schak -rw-rw----    1 root     mail         1500 Dec 26 05:48 root
1104101953 M * Bertl and btw, what does grep mail /etc/group return?
1104101958 M * ndim Bertl: If no one has ever used any vserver specific userspace tools in a guest system, then the -guest package is obsolete, of course.
1104101979 M * Schak base:/var/mail# grep mail /etc/group
1104101979 M * Schak mail:x:8:
1104101983 J * DuckMaster ~Duck@dyn-83-157-168-176.ppp.tiscali.fr
1104101988 M * Bertl welcome DuckMaster!
1104101988 M * ndim I just got the impression there were such tools when I tried to find out how util-vserver alpha work.
1104102006 M * Doener ndim: actually most if not all tools aren't allowed to be used from within the vserver
1104102055 M * ndim Doener: Very good. Most or all? :)
1104102065 M * Schak wow... i created with adduser a new user... and this user cant login, too
1104102123 M * Bertl some special security?
1104102140 M * Doener things like v_inetd are meant be used in the host to help services that can't bind to a specific ip themselves (or for lazy admins, your choice)
1104102155 M * Bertl Schak: apropos security, you have not accidentially compiled the capabilities as module and not loaded it?
1104102185 M * Doener ndim: none are meant to be used from within the vserver, and i don't think any have any effect when used from within
1104102195 M * Schak module support is completely off, everything i need is directly compiled into the kernel
1104102210 M * Doener ndim: f.e. the vserver syscall is protected by CAP_CONTEXT, which no vserver should have
1104102244 M * Schak [00:02:39] <Bertl> some special security? <-- its the default woodie install from debootstrap... nothing specific handmade stuff
1104102284 M * Bertl Schak: you are logged on via ssh or 'enter' ?
1104102296 M * Schak i tried both
1104102313 M * Schak via root works both... via user neighter nor
1104102314 M * ndim Doener: OK. So the use case for v_inetd is to have a vserver kernel and run inetd on a specific IP in the host system's context?
1104102356 M * Bertl Schak: okay, logon via ssh (as root) restart the syslog service from there, and tail -f the output /var/log/messages (or whatever)
1104102374 M * Bertl then try to logon via ssh as user x/y
1104102383 M * Doener ndim: exactly, if it would listen on 0.0.0.0 it would prevent the vservers from using the same ports on 'their' ip addresses. But I'd prefer to change inetd's configuration instead ;)
1104102401 M * ndim Oh, I see.
1104102415 Q * duckx Ping timeout: 480 seconds
1104102417 M * Bertl you basically can use those wrappers for all services except sshd
1104102420 M * Doener btw using v_sshd is a very bad idea ;)
1104102443 M * Bertl (sshd is special in that way that you usually want to administer your vservers over ssh too)
1104102444 M * Schak Dec 26 23:54:15 base <someone>: Hello folks
1104102444 M * Schak Dec 27 00:06:32 base exiting on signal 15
1104102444 M * Schak Dec 27 00:06:33 base syslogd 1.4.1#10: restart.
1104102494 M * ndim Well, sshd can be configured to listen to a specific IP anyway.
1104102519 M * Bertl Schak: the sshd logon isn't even reported?
1104102538 M * Doener yeah, but lazy folks may tend to use v_ssh no matter what ;)
1104102556 M * Bertl imho v_sshd should be removed ...
1104102590 M * Schak i cant login with a normal user... same error... connection closed by foreign host
1104102607 M * Schak [00:09:15] <Bertl> Schak: the sshd logon isn't even reported? <-- yes, its in /var/log/auth.log
1104102635 M * Bertl ah, okay, and the sshd doesn't report anything?
1104102677 M * Schak a paste the last line from auth.log into the query window
1104102919 M * Bertl let's try to run the sshd with -d  (on a different port)
1104102948 M * Bertl (e.g. sshd -d -p 9999)
1104103035 M * Doener 	inode->i_gid = INOXID_GID(XID_TAG(inode), uid, gid);
1104103037 M * Schak ok
1104103042 M * Doener +#define XIDINO_GID(tag, gid, xid)	(gid)
1104103077 M * Doener that becomes: inode->i_gid = uid , right?
1104103151 M * Schak login_get_lastlog: Cannot find account for uid 1000
1104103176 M * Bertl Doener: the way you present it here, yes
1104103186 M * Schak base:/# grep 1000 /etc/passwd
1104103186 M * Schak pzystorm:x:1000:1000:,,,:/home/pzystorm:/bin/bash
1104103217 M * Doener hm, did i mess up something?
1104103251 M * Bertl sec, I'm verifying (which tagging scheme did you assume?)
1104103261 M * Schak 24bit
1104103261 M * Doener oops... INOXID vs. XIDINO...
1104103312 M * Bertl Schak: how good is your network connection, and how large is the vserver?
1104103343 M * Bertl (or is it possible to get access to that vserver somehow?)
1104103348 M * Schak 100mbit... "large"?... hd size?... amount of vserver?
1104103351 M * Schak yes of course
1104103358 M * Schak query
1104103763 M * Bertl okay, problem solved ... please Doener, explain ... ;)
1104103787 M * Bertl (the honor is yours, as you spotted it first :P)
1104103800 M * Doener all permissions for the vserver's root directory were revoked, thus nobody except root had any right to access any files
1104103835 M * Bertl maybe also a good idea for a FAQ if not already there?
1104103874 M * Doener those permissions were meant to be set for the /vservers directory (plus chattr +t) with the 2.4 patches to protect from chroot escapes, but this is no longer necessary as we now have a special flag to achieve the same
1104103912 M * Bertl (the barrier flag, and it isn't even neccessary for namespace based servers anymore)
1104104090 M * Bertl Schak: exim now working as expected?
1104104361 M * Schak yes... problem no2 was, that i had running qmail on the host listening at 0.0.0.0 on port 25... so exim on the vserver cant listen... after stopping qmail, i restarted exim... and now everything works
1104104371 M * Bertl Doener: what do you think of the split, are the logical units sound for you?
1104104391 M * Bertl Schak: excellent ...
1104104584 M * Doener i don't know yet if it is wise to put the includes into a seperate patch... 
1104104612 M * Bertl well, I did that to basically get rid of them ...
1104104637 M * Bertl and it already reduced the number of included files)
1104104684 M * Doener ok, if you can track them easier that way, it's fine with me. I'm bad at keeping track of them anyways ;)
1104104742 M * Bertl every now and then I remove one (i.e comment them out) and see if it compiles ;)
1104104748 M * Doener hehe
1104104793 M * Bertl but splitting out the cmdef stuff was a great improvement regarding kernel rebuilds ...
1104105027 M * Doener hm?
1104105092 M * Bertl well, in the beginning I had the init/exit/proc stuff in the definition files, which resulted in a complete kernel recompile everytime I changed something
1104105101 A * ndim is feeling weak and goes fetch more Kaesspaetzle
1104105407 M * Bertl hmm, a snack .. good idea ...