1137974603 M * Bertl well, to me it always seemed like, nobody really understood it, but as it 'worked' with pid, it was just fine :)
1137974649 M * azazel 'night!
1137974668 M * Bertl azazel: night!
1137974680 Q * azazel Quit: Client exiting
1137974692 M * ebiederm Bertl: I think that is half the problem understanding the model in the kernel is a challenge.
1137974719 M * ebiederm Looking at what the kernel does It just occured to me that pid really is a pid.
1137974756 M * ebiederm And that tgid is a process group above pids but below normal process grous.
1137974786 M * ebiederm Because kill(pid, ...) on any thread idea actually works.
1137974816 M * ebiederm It is still confusing because it does not exactly match the user space concept.
1137974825 M * Bertl yes, indeed
1137974873 M * ebiederm And of course get_pid has been modfied to be get_tgid and get_tid is really get_pid!
1137975170 J * shedi ~siggi@inferno.lhi.is
1137975266 J * Smutje_ ~Smutje@xdsl-87-78-2-172.netcologne.de
1137975379 Q * Smutje Ping timeout: 480 seconds
1137975718 M * ebiederm What is the difference to a tty of pgrp 0 versus pgrp -1?
1137975864 M * Bertl no idea ... at,
1137975864 M * Bertl *atm
1137975906 M * ebiederm I don't either but the code treats the two cases just enough different that there may be a difference.
1137977121 Q * jpacheco Quit: BitchX-1.1-final -- just do it.
1137978478 M * ebiederm Hmm. User space is allowed to set tty->pgrp to 0 but not -1....
1137978651 M * Bertl maybe when the pgrp was lost?
1137978673 M * ebiederm Or cleared.
1137978707 M * ebiederm I will get it.  Right now I'm handling the obvious cases... and making notes as I go along.
1137978832 M * Aiken Bertl I already have one problem with 2.6.16-rc1
1137978849 M * Aiken everytime I run make CROSS_COMPILE=alphaev5-unknown-linux-gnu- ARCH=alpha menuconfig  /dev/null gets deleted :(
1137978885 M * Bertl wow, cool!
1137978902 M * Cru heh
1137978963 M * Bertl I guess (hope) Sam has missed something ...
1137979010 M * Bertl Aiken: could you do that with -v once again (and upload the output)?
1137979224 M * Aiken a lower case -v? just returns version number a std gnu make 3.80
1137979271 M * Aiken does not even have to be cross, a normal make mrproper does it as well
1137979611 M * Aiken in menuconfig, instead of the options being centered they are to the left, 2.6.15 is behaving on my system
1137979741 M * Bertl ah, hmm, confusingly it's -d 
1137980048 M * derjohn n8 folks ...
1137980057 M * Bertl good nigth derjohn!
1137980062 M * Aiken if I make -d mrproper &> /tmp/delme  or make -d whatever /dev/null does not show up
1137980073 M * Aiken but it does go and return as a normal file
1137980200 M * Aiken the top few lines from make mrproper -d &> /tmp/delme    http://pastebin.com/518412
1137980219 M * Aiken wondering about these lines  Got a SIGCHLD; 1 unreaped children.
1137980271 M * Aiken 2.6.15 has them as well
1137980388 M * Aiken you sure you all of one of the -d files? one is over 322,000 lines and 2 meg in size
1137980420 M * Aiken sorry, not 2meg - 20meg
1137980625 M * Aiken just found this http://bugzilla.kernel.org/show_bug.cgi?id=5916
1137981267 J * marl ~matt@84.92.193.226
1137981281 M * Bertl welcome marl!
1137981298 M * Bertl Aiken: yeah, well, I suspected a kernel build system issue ...
1137981326 M * marl i there
1137981338 M * marl  s/i/hi/
1137981342 M * Aiken means I am a bit reluctant to start putting 2.6.16 source on my machines
1137981399 M * Bertl Aiken: that's fine, it will have to wait for 2.6.16-rc2 ...
1137981467 M * marl quick question, is it posible to get vserver to create full interfaces without the hardware being there? e.g. vserver host has eth0 and eth1 on it, both connected to networks, i want to create 2 vservers one with an interface 'eth2' and one with an interface 'eth3', is this posible? wanting to do it this way to simplfy firewalling and seperating vservers from other parts of the network.
1137981542 M * marl but this with just 2 hardware interfaces in the host system, and use firewall rules ONLY to forward packetrs between the interface, like a virtual network running within the host system, if that makes any sence!!!!
1137981595 Q * pusling Remote host closed the connection
1137981615 J * pusling pusling@195.215.29.124
1137981741 M * Bertl marl: no, but you can use dummy0 for the guests
1137981764 M * Bertl marl: but keep in mind that networking happens on the host, so nothing will be sent to dummy0 anyway
1137981790 M * marl ah, :(
1137981797 M * Bertl marl: so nothing is 'forwarded' between guests and hosts
1137981813 M * Bertl marl: nevertheless you can use all firewalling tricks you have on the host
1137981839 M * marl thats th eproblem im having, trying to get the guest isolated from the host as much as posible
1137981945 M * marl i setup a host with an ip of 192.168.0.36 with sshd set to 192.168.0.36:22 and a guest with an ip of 192.168.0.3, dispite the sshd in the host being tied to 192.168.0.36 it still picked up requests for ssh 192.168.0.3 :(
1137982019 M * Bertl unlikely
1137982050 M * Bertl if the sshd would have been tied to 192.168.0.36 it would not pick up requests for 192.168.0.3
1137982070 Q * vrwttnmtu Quit: Leaving
1137982077 M * Bertl more likely it is bound to 0.0.0.0, so it picks up everything
1137982111 M * Bertl check/add a Listen directive in your config and make sure to restart the daemon
1137982118 J * mire ~mire@62.108.127.196
1137982132 M * mire hello
1137982134 M * Bertl welcome mire!
1137982198 M * mire is it possible to install fc4 on my debian? I tried something like vserver fc4 build -n fc4 -m yum --hostname=fc4.beast.kom --interface test0=eth0:192.168.0.26/24 --initstyle sysv -- -d fc4 and got an error mount: mount point /etc/rpm does not exis
1137982226 M * Bertl well, it _is_ possible, but debian makes it very hard ...
1137982245 M * Bertl the key issue is getting yum/apt-rpm/rpm working as expected
1137982283 M * Bertl usually it's much simpler to use any rpm based host to install the fc4 guest (with util-vserver) and then copy it over to the debian machine
1137982353 M * mire ok, I already have a filesystem with rh9, can I copy that under /var/lib/vservers?
1137982398 M * Bertl sure, the correct/simple sequence here is:
1137982414 M * Bertl - create a skeleton guest with the complete config (--interfaces and such)
1137982436 M * Bertl - copy over everything except the /dev directory into the newly created guest dir
1137982442 M * Bertl - start the guest
1137982466 J * pusling_ pusling@195.215.29.124
1137982576 M * marl grep List /etc/ssh/sshd_config |grep -v '^#'
1137982576 M * marl ListenAddress 192.168.0.36
1137982590 Q * pusling Ping timeout: 480 seconds
1137982593 M * mire Bertl, thank you
1137982641 M * Bertl mire: you're welcome!
1137982664 M * Bertl mire: btw, you might want to cleanup the guest (i.e. disable hardware related scripts and such)
1137982720 M * marl ok just confirm on the host that ssh is bound to *:* :(
1137982777 Q * pusling_ Remote host closed the connection
1137982786 J * pusling pusling@195.215.29.124
1137983625 J * Aiken_ ~james@tooax8-126.dialup.optusnet.com.au
1137983752 M * marl ok, just for information , running /etc/init.d/sshd restart, booted me out of my ssh sesion (making me think sshd was restarting) but DIDNT re-read the config file properly, didnt pick up the change untill host rebooted :(
1137983795 M * marl now working, thanks Bertl , it was jsut my own stupidity :(
1137983909 M * mire hey, what do you think of using debian.marlow.dk images?
1137983974 Q * Aiken Ping timeout: 480 seconds
1137984164 M * Bertl marl: np, happens ...
1137984176 M * Bertl mire: well, if they suit your needs, go ahead
1137984271 M * marl ok heres another question :) apart from using ssh to connect and run commands, is there a way to comunicate with the host system?
1137984328 M * marl i know it kinda goes against the whole vserver/chroot idea, but for VERY limited comunication may be handy sometimes
1137984354 M * Bertl sockets, pipes, files
1137984370 M * Bertl (given there is something running on the host)
1137984399 M * marl ok thanks, had started looking at sockets, and thought id check before re-inventing the wheel :)
1137984489 M * marl ok, one other nd ill out of your hair for the night (hopefully!!!!) is it posible to run a guest with its chroot set to read-only? (without setting the entire vserver directory to read-only in the host)?
1137984581 M * Bertl yes, with a bind mount, given that you are using a devel version or have BME patches applied
1137984595 M * Bertl in this case, you can simply do mount --bind -o ro
1137984637 M * marl would that be on a guest by guest basis? if so which file would it go into?
1137984658 M * Bertl you could do that with the fstab
1137984726 M * marl i didnt think fstab was used in the guest systems, unless u were creating a temp memory fs
1137984770 M * Bertl there is one for each guest (in the tree based config)
1137984858 M * marl ah, i see, ive got proc/tempfs/devfs, in there at hte moment, is there any docs on how to setup options in the file?
1137984920 M * Bertl it's like in fstab, just that the 'device' is in the host namespace, while the mount point in the guest
1137985033 M * marl so if /dev/hda6 was mounted /vserver in the host, the fstab for the guest would be /vserver/guest1 / ? or would it be /dev/hda6 / ?
1137985204 M * Bertl for the bind mount you need pathes
1137985222 M * marl ah, thanks :)
1137986107 M * mire I have a problem with routing... can't reach the internet
1137986114 M * mire local network works fine 
1137986247 M * marl mire do u have dns setup within the vserver?
1137986257 M * marl (i take it this is whenur in the vserver?)
1137986275 M * mire marl, this is not dns problem
1137986288 M * mire its a routing problem (I can't ping)
1137986293 M * mire an IP
1137986298 M * marl ah ok, sorry
1137986322 M * marl is your vserver machine connected directly to the net?
1137986333 M * mire yep, I'm typing this from host box
1137986362 M * marl is the host used as a router for the rest of your network, or is it just a standalone?
1137986395 M * mire I have just this box
1137986441 M * _Roey hey all
1137986443 M * _Roey hello Bertl
1137986516 M * marl mire, i know it sounds daft, but have u enabled routing for the host?
1137986557 Q * mkhl Quit: 
1137986570 M * mire I think so...
1137986590 M * mire let me check it out
1137986593 M * marl can u remember how u enabled it?
1137986696 M * mire echo 1 > /proc/sys/net/ipv4/ip_forward
1137986725 M * marl ok, sounds right so far :) can u ping th eip of the host?
1137986761 M * mire yep, local network works ok
1137986765 M * mire but internet won't
1137987015 M * marl i seem to remember hittin gthis myself, but i cant remember how to solve it, its 3.30am here and my head is bouncing off the keyboard every coupl eof mins :(
1137987027 M * marl do u have iptables on the host?
1137987056 M * mire yep
1137987075 M * marl have u defined any iptables stuff yet?
1137987121 M * daniel_hozac mire: iptables -t nat -A POSTROUTING -o <external interface> -j SNAT --to <external IP address>
1137987166 M * marl thanks daniel_hozac , i was just running through my iptables stuff trying to find that one :)
1137987185 M * marl ok folks im headin gto bed, speek u lot later
1137987202 M * mire marl, n8
1137987277 M * mire daniel_hozac, works!!!
1137987296 M * mire daniel_hozac, you have a drink from me :)
1137987308 M * daniel_hozac heh, thanks.
1137988097 M * mire anyone know how do I remove an image?
1137988134 M * Bertl rm -rf ?
1137988147 M * Bertl but make sure to stop the guest first ...
1137988158 Q * dlippolt Quit: Leaving
1137988188 M * mire yeah did that
1137988192 M * ebiederm I have just observed a FAQ.
1137988205 M * Bertl ebiederm: yes :)
1137988348 M * mire sorry about the faq q :)
1137988376 M * Bertl mire: np ... we love to help ...
1137988387 M * ebiederm I was laughing because I saw the same question and Bertl give the exact same problem just a couple of days ago.
1137988406 M * mire btw its cool just rm -fr ;)
1137988412 M * mire I like that
1137988414 M * ebiederm I had never observed a FAQ quite like that before.
1137988492 M * Bertl well, the interesting part is, previously I just said "remove it (with rm -rf or so)"
1137988524 M * Bertl and usually after a few minutes I got: "I can't remove it, it says it's busy"
1137988557 M * Bertl which then results in "did you stop it?" .. "no!" .. "well, bad luck!" :)
1137988580 M * ebiederm What happens if you don't stop it first?
1137988584 M * Bertl and a longer explanation how to kill of the processes, enter the namespace, unmount the stuff ...
1137988604 M * ebiederm Because the automatic scripts can't kill it!  Got it.
1137988613 M * Bertl and then remove the remaining parts ...
1137988638 M * Bertl actually I would prefer a 'vserver name remove' or delete
1137988656 M * Bertl but for some reason, enrico (the guy doing the tools) does not like that
1137988679 M * mire so what does enrico prefer?
1137988698 M * Bertl obviously, 'vserver name stop' and rm -rf :)
1137988782 M * daniel_hozac Bertl: hmm, why doesn't he like it?
1137988785 M * Bertl but well, it should be trivial to add ...
1137988799 M * Bertl daniel_hozac: no idea, I suggested it several times and got no feedback
1137988859 M * Bertl daniel_hozac: similar for the 'copy' method ... I suggested to use the same interface as for 'build', just to copy a template (or existing guest) over to the dir .. IIRC, I got an reply to that 'request' stating that this is more complicated ...
1137988884 M * Bertl but maybe that was added in 0.30.210, haven't checked yet
1137988897 M * Bertl btw, any issues with 0.30.210?
1137988931 M * daniel_hozac no, nothing like that in 0.30.210.
1137988961 M * Bertl time for super-daniel the patch-man?
1137988965 M * daniel_hozac not yet, it seems to work ok.
1137988988 M * daniel_hozac hehe, i'll see what i can do.
1137988994 M * Bertl it seems to break with 1.2.10 kernels ...
1137989013 M * daniel_hozac yes, by default.
1137989017 M * Bertl first I had to enable the compat interfaces, which worked quite fine
1137989028 M * Bertl interestingly the legacy interfaces did not even compile
1137989050 M * Bertl then I got this issue: https://savannah.nongnu.org/bugs/index.php?func=detailitem&item_id=15508
1137989146 M * daniel_hozac that's odd.
1137989183 M * Bertl have to build them for my testsystem and verify ...
1137989312 M * Bertl I guess the mips/hppa update didn't make it into 0.30.210 either
1137989334 M * daniel_hozac the chain of commands didn't change, so i don't see what would make it EPERM...
1137989778 M * daniel_hozac what APIs did you enable? v11,v13,net,fscompat,compat (i.e. old default)?
1137989986 M * Bertl sec
1137990054 M * Bertl everything mentioned on the configure, except for the legacy
1137990123 M * Bertl --enable-apis=compat,v11,fscompat,v13obs,v13,net
1137990141 M * daniel_hozac ah.
1137990158 M * Bertl to my surprise the configure also mentioned two others, but giving them was rejected
1137990170 M * daniel_hozac oldproc and olduts?
1137990174 M * Bertl yep
1137990278 M * daniel_hozac v11, legacy and compat imply them, apparently. configure should still accept them though...
1137990311 M * Bertl well, configure bails out with some strange message, try for yourself :)
1137990324 M * daniel_hozac yeah, oldproc is not a supported API.
1137990429 M * Aiken_ Bertl trying to build a 2.6.16-rc1-git4-vs2.1.0.6.1 kernel
1137990447 M * Bertl wow, patch did apply?
1137990483 M * Aiken_ 3 x .rej to deal with
1137990494 M * Aiken_ 2.6.16-rc1 won't even compile for me
1137990538 M * Bertl for some reason I kept my /dev/null on my alpha
1137990540 M * Aiken_ even had to resort to compiling the kernel as a normal user so not to kill off /dev/null
1137990574 M * Aiken_ there is a thread about this lkml, was getting the impression some do and some don't
1137990594 M * Bertl yeah, seems to depend on a special mix of tools and setup
1137990671 M * Aiken_ this look famillar?
1137990674 M * Aiken_ fs/open.c: In function `sys_fchownat':
1137990674 M * Aiken_ fs/open.c:757: warning: passing arg 2 of `chown_common' makes pointer from integer without a cast
1137990674 M * Aiken_ fs/open.c:757: error: too few arguments to function `chown_common'
1137990711 M * Bertl yeah, could be an issue with the bme patches
1137990731 M * Bertl could you try to compile the following:
1137990768 M * Aiken_ I am currently trying a std 16-rc1 for alpha to make sure that compiles
1137990813 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.16-rc1-bme0.06.2.diff (just the bme patch)
1137990835 M * Bertl if vanilla works, I'd be interested if this one fails ...
1137990902 M * Aiken_ vanilla gets past fs/open.c
1137990957 M * Aiken_ should I do that bme patch with -R ?
1137990971 M * Bertl hmm, no, just ontop of vanilla
1137990983 M * Bertl sec, checking
1137990986 M * Aiken_ ok
1137991012 M * Bertl does it look like it wants -R ontop of vanilla?
1137991027 M * Bertl (would be an indication that some parts were actually included)
1137991042 A * Bertl wanders off reading the changelog
1137991068 M * Aiken_ no, that was on top of 2.1.whatever
1137991076 M * Bertl ah, okay :)
1137991152 M * Aiken_ just redoing a couple of tests, have 4 trees I am playing with
1137991167 M * Bertl you lucky guy you :)
1137991388 M * Bertl wow, uml gets an overhaul ...
1137991419 M * Aiken_ trying that bme patch on rc1-git4
1137991436 M * Bertl okay, now it's getting interresting ...
1137991464 M * Aiken_ the scsi driver I want won't even compile in rc1 so rc1 is of no use to me
1137991501 M * Bertl does this apply to rc1-git4 too?
1137991519 M * Aiken_ fs/open.c -> same error as above
1137991539 M * Bertl okay, I'll look into it .. give me a few minutes
1137991728 M * Aiken_ many Hunk lines but no rejects http://pastebin.com/518594
1137992008 M * Aiken_ I changed the line to  error = chown_common(nd.dentry, nd.mnt, user, group);
1137992046 M * Aiken_ based on all the other calls to chown_common
1137992162 M * Bertl yep, the at syscall is new
1137992171 J * undefined ~undefined@adsl-68-93-109-94.dsl.rcsntx.swbell.net
1137992284 M * mire I have an error with the rpm
1137992295 M * mire [root@vserver:rh9 init.d]rpm -qa
1137992295 M * mire rpmdb: unable to join the environment
1137992295 M * mire error: db4 error(11) from dbenv->open: Resource temporarily unavailable
1137992295 M * mire error: cannot open Packages index using db3 - Resource temporarily unavailable (11)
1137992295 M * mire error: cannot open Packages database in /var/lib/rpm
1137992297 M * mire no packages
1137992321 M * Bertl inside a guest, or on the host?
1137992331 M * Bertl if guest, how was it installed?
1137992346 M * mire guest
1137992360 M * mire I took an image from marlow
1137992376 M * Bertl well, maybe no rpm database is inside (or outside)
1137992385 M * mire but if I su - someuser
1137992388 M * mire rpm -qa works
1137992412 M * Bertl how do you access the guest?
1137992426 M * mire vserver rh9 enter
1137992437 M * Bertl okay, use ssh to enter it as root
1137992448 M * Bertl the 'enter' takes the current environment with you
1137992457 M * Bertl (which is probably not suitable for the guest)
1137992475 M * mire I used your advice and copied a custom redhat image after
1137992485 M * mire so now I only have marlow's /dev
1137992494 M * mire just a sec.
1137992497 M * Bertl hmm ... 
1137992501 A * Bertl scrolls up ...
1137992527 M * mire Bertl - create a skeleton guest with the complete config (--interfaces and such)
1137992528 M * mire Bertl - copy over everything except the /dev directory into the newly created guest dir
1137992528 M * mire Bertl - start the guest
1137992529 M * Bertl 03:13 < Bertl> - copy over everything except the /dev directory into the newly  created guest dir
1137992533 M * mire :)
1137992559 M * Bertl so .. how could it happen that you have marlow's dev then? :)
1137992597 M * mire I installed marlows image
1137992605 M * mire then deleted everything xcept for /dev
1137992610 M * mire and installed my own image
1137992614 M * mire (copied)
1137992622 M * Bertl now I'm confused ...
1137992631 M * Bertl what's the point in doing so?
1137992634 M * mire rpm behaves the same in both configs
1137992653 M * mire well rh9 install doesn't work on debian
1137992663 M * mire so to get it working I installed marlows image
1137992675 M * mire then to get my rh9 working I copy the stuff
1137992686 M * mire per your advice
1137992689 M * mire and it works
1137992692 M * Bertl okay
1137992697 M * mire but... rpm -qa behaves the same
1137992705 M * mire in marlows image and now in my image
1137992711 M * Bertl did you try to ssh into the guest?
1137992715 M * mire I'm doing sshd setup, just a sec. :)
1137992723 M * Bertl (instead of using the enter)
1137992923 M * mire [root@vserver:rh9 init.d]/etc/init.d/sshd start
1137992923 M * mire Starting sshd:                                             [  OK  ]
1137992929 M * mire but ssh is not running :/
1137992944 M * Bertl well, because despite the message is was not started
1137992964 M * Bertl why? simple: because sshd is already running on the host, and did bind to 0.0.0.0
1137992976 M * mire I made a typing error :)
1137992979 M * mire just a sec.
1137993030 M * Bertl if you land on the host instead of the guest, the above is valid
1137993058 M * mire ok, I got in
1137993073 M * mire gives me the same error
1137993095 M * Bertl very strange ...
1137993105 M * mire but if I do su - something
1137993108 M * mire it works
1137993113 M * Bertl what about su - root
1137993120 M * mire with an occasional error
1137993122 M * mire error: rpmdbNextIterator: skipping h#      67 Header V3 DSA signature: BAD, key ID db42a60e
1137993183 M * mire [root@vserver:rh9 root]rpm --rebuilddb
1137993183 M * mire rpmdb: unable to join the environment
1137993183 M * mire error: db4 error(11) from dbenv->open: Resource temporarily unavailable
1137993183 M * mire error: cannot open Packages index
1137993212 M * Bertl does 'su - root' inside the guest help too?
1137993257 M * mire no, it gives the same error
1137993282 M * mire http://www.redhat.com/archives/rpm-list/2003-March/msg00057.html
1137993291 M * mire I see some other people had the same problem
1137993321 M * daniel_hozac RHL9 problem, it seems.
1137993326 M * daniel_hozac http://www.fedora.us/pipermail/fedora-devel/2003-April/001088.html has some workarounds.
1137993338 M * Bertl mire: search for __ files
1137993354 M * Bertl sometimes those temporary files keep rpm from working
1137993387 M * mire just a sec.
1137993428 M * mire I got __db.001
1137993432 M * mire 0 bytes
1137993435 M * mire will delete it
1137993459 M * mire ok, when I did rpm -qa it made again that same file
1137993495 M * mire LD_ASSUME_KERNEL=2.2.5 rpm -qa helped
1137993520 M * mire it works as when I am an ordinary user
1137993534 M * mire with some mistakes as I pasted earlier
1137993535 M * Bertl that means the rpm is confused by some NPTL issues ...
1137993545 M * Bertl what kernel do you use on the host?
1137993562 M * mire 2.6
1137993573 M * Bertl which one?
1137993579 M * mire Linux zwer 2.6.14.3-vs2.0.1-vserver-1
1137993651 M * Bertl could you 'strace -fF -o rpm.trace'  that for me, both with and without the LD_ASSUME_KERNEL=2.2.5 ?
1137993659 M * mire sure
1137993669 M * Bertl use different .trace files and either upload or dcc them to me
1137994152 M * mire ok, I made them, just a sec.
1137994294 M * mire http://osadmin.com/linux/strace/
1137994321 M * Aiken_ Bertl unhandled unaligned exception :(
1137994333 M * Aiken_ solid lockup
1137994333 M * Bertl hmm, where?
1137994360 M * Aiken_ probably starting a guest, probably
1137994361 M * Aiken_ maybe
1137994375 M * Aiken_ nothing definate on screen and nothing from netconsole
1137994399 M * Bertl shouldn't the alpha kernel fix up the broken alignments on the fly?
1137994414 M * Aiken_ usually
1137994518 M * Bertl mire: what filesystem do you use for the guest?
1137994523 M * mire reiserfs
1137994537 M * Bertl could you try with ext2/3 too?
1137994571 M * Bertl (simple way would be to create a loopback, copy the guest over and mount that)
1137994582 M * mire yeah, could do that
1137995210 M * Aiken_ this is not good http://pastebin.com/518663
1137995294 M * mire Bertl, mount --bind?
1137995314 M * mire or is symlink enough?
1137995315 M * Bertl or just mount, but make sure to stop the guest first :)
1137995348 M * Bertl mount /dev/loop0 /vservers/guest (or wherever your guests are)
1137995373 M * mire ok
1137995431 M * Bertl Aiken_: hmm, somehow looks to me like this happens with the (kernel) init task ...
1137995456 M * Bertl Aiken_: could you try to lookup fffffc000032a608 with gdb or (the patched) addr2line?
1137995469 M * Aiken_ non of the guests were started
1137995484 M * Bertl yes, it happens right before init is started
1137995525 M * Bertl would also be interesting to know if 2.6.16-rc1-git4 works
1137995555 M * Aiken_ last 4 lines from a good boot
1137995559 M * Aiken_ Jan 19 15:41:41 pebbles input: AT Translated Set 2 keyboard as /class/input/input0 
1137995559 M * Aiken_ Jan 19 15:41:41 pebbles input: ImPS/2 Generic Wheel Mouse as /class/input/input1 
1137995559 M * Aiken_ Jan 19 15:41:41 pebbles Adding 288464k swap on /dev/sda4.  Priority:2 extents:1 across:288464k 
1137995559 M * Aiken_ Jan 19 15:41:41 pebbles EXT3 FS on sda2, 
1137995559 M * Aiken_ Jan 19 15:41:41 pebbles internal journal 
1137995658 M * Bertl but it is definitely vserver related
1137995871 M * mire /dev/loop0            4.0G  3.4G  408M  90% /var/lib/vservers
1137995938 M * Bertl Aiken_: hmm, IIRC seccomp is now disabled by default
1137995964 M * Bertl this might be far-fetched, but, what if the seccomp_t struct
1137995973 M * Bertl (which is an int, basically)
1137995992 M * Aiken_ (root@pebbles) addr2line -e vmlinux fffffc000032a608
1137995992 M * Aiken_ include/asm/system.h:406
1137996005 M * Bertl messes up the task_struct (or something else does) so that the vx_info is misaligned for the init process
1137996024 M * Bertl no idea why the kernel doesn't fix that up, but ...
1137996069 M * Bertl yep, that's the xchg from __enter_vx_admin()
1137996071 M * mire Bertl, I get the same error
1137996085 M * Aiken_ this is the function http://pastebin.com/518674
1137996113 M * Bertl mire: okay, seems it is kernel + rpm related ...
1137996129 M * Bertl mire: I don't see anything going wrong in the traces
1137996129 M * mire can I fix it?
1137996145 M * mire maybe if I install fedora core 2 it will work
1137996145 M * Bertl mire: well, yes, probably by changing the rpm or kernel
1137996399 M * Bertl Aiken_: let's try to put some __attribute__((__aligned__(8)))
1137996422 M * Bertl right after the vx_info and nx_info structs in task_struct
1137996456 M * Bertl include/linux/sched.h (699)
1137996487 M * Bertl and a general __attribute__((__aligned__(16))) at the end of the entire struct
1137996604 M * Bertl Aiken_: do you need a patch for that?
1137996639 M * Aiken_ we shall see
1137996646 M * Aiken_ I found those 2 bits
1137996677 M * Bertl struct vx_info *vx_info __attribute__((__aligned__(8)));
1137996702 M * Bertl (should do the trick)
1137996708 M * Aiken_ that is what I did
1137996718 M * Bertl and at the end of the struct:
1137996731 M * Bertl } __attribute__((__aligned__(16)));
1137996736 M * Bertl just to make sure ...
1137996740 M * Aiken_ already compiling
1137996745 M * Bertl excellent!
1137996815 M * Hollow Aiken_: did you upgrade to vserver-utils-1.0.3 too?
1137996820 M * Hollow morning btw ;)
1137996842 M * Aiken_ Hollow no
1137996849 M * Aiken_ I had tried 1.0.2 of each
1137996854 M * Hollow the bad address thing was fixed there
1137996894 M * Hollow (any many many more bugs)
1137996958 M * Bertl Aiken_: ah, it seems that unaligned access by the kernel is not fixed up
1137997009 M * Aiken_ the 3 attribute lines did not work for you?
1137997038 M * Bertl no, I meant the fact that the unaligned access was not handled
1137997050 M * Bertl and instead produced the kernel oops
1137997111 M * Aiken_ just realised, that oops was the 1st serious use of netconsole here
1137997219 M * Bertl could you use the addr2line (or whatever you used) once again for the offending address (fffffc00005daeb5)
1137997266 M * Bertl Aiken_: the output of the scsi subsystem is indeed interesting (the one char per message part)
1137997307 M * Aiken_ (root@pebbles) addr2line -e vmlinux fffffc00005daeb5
1137997307 M * Aiken_ ??:0
1137997367 M * Aiken_ looking at System.map I get
1137997369 M * Aiken_ fffffc00005d9bf5 r __func__.1
1137997369 M * Aiken_ fffffc00006459f8 r __pci_fixup_PCI_VENDOR_ID_INTELPCI_DEVICE_ID_INTEL_PXHD_0quirk_pcie_pxh
1137997430 M * mire Bertl, fedora works! :)
1137997441 M * Aiken_ the other address on that line is fffffc000032a608    include/asm/system.h:406
1137997584 M * Bertl mire: excellent ..
1137997607 M * Bertl Aiken_: yeah, it's probably the init structure
1137997698 M * Aiken_ include/asm/system.h:406 is in http://pastebin.com/518674
1137997741 M * Aiken_ 406 is the __asm__ __volatile__( line
1137998090 M * Aiken_ Linux pebbles.bedrock 2.6.16-rc1-git4-vs2.1.0.6.1 #2 Mon Jan 23 16:27:15 EST 2006 alpha Unknown Alcor GNU/Linux
1137998113 M * Bertl excellent!
1137998122 M * Aiken_ those 3 align attributes did something good
1137998138 M * Bertl I would assume the first one was sufficient
1137998160 M * Bertl I can't believe that the others are actually required
1137998193 M * Bertl nx_info is not xchanged, and a misalignment for the entire struct would probably cause havok elsewhere ...
1137998209 M * Aiken_ (root@hoppy) uname -a
1137998210 M * Aiken_ Linux hoppy.bedrock 2.6.16-rc1-git4-vs2.1.0.6.1 #2 Mon Jan 23 16:27:15 EST 2006 alpha Unknown Alcor GNU/Linux
1137998214 M * Aiken_ that is a cow guest
1137998397 M * Bertl even better, so the cow link break is supposedly working to, right?
1137998421 M * ebiederm s/to/too/ ?
1137998422 M * Aiken_ my cow-test script from the inital testing shows the link breaking working
1137998502 M * Aiken_ just deleted many files in a new cow guest, they are gone and the files in the master copy are still there
1137998570 M * Bertl ebiederm: yes, tx :)
1137998645 M * Bertl Aiken_: okay, thanks a lot for testing, I will rework the patches and figure something for the next release (to completely eliminate the alignment issues for alpha)
1137998685 M * Bertl I'm off to bed now .. so have fun everyone! and cya later ...
1137998688 M * Aiken_ it is silly how long it took to get to this point
1137998713 M * Aiken_ bye bye
1137998716 M * Bertl cya
1137998719 M * ebiederm Bertl: cya later.  I should have an interesting proof of concept working sometime tommorrow :)
1137998727 N * Bertl Bertl_zZ
1137998739 M * Bertl_zZ ebiederm: great! and good night!
1137998954 J * JochenA ~jochen@12-216-244-216.client.mchsi.com
1137999499 Q * pusling Read error: Connection reset by peer
1137999502 J * pusling_ pusling@195.215.29.124
1137999945 Q * pusling_ Read error: Connection reset by peer
1137999947 J * pusling pusling@195.215.29.124
1138000859 Q * JochenA Quit: Client exiting
1138001912 J * Aiken__ ~james@tooax6-166.dialup.optusnet.com.au
1138001981 Q * bubulak Ping timeout: 480 seconds
1138002183 Q * klap Remote host closed the connection
1138002229 Q * Aiken_ Ping timeout: 480 seconds
1138002554 J * bubulak ~bubulak@cicka.wnet.sk
1138003596 Q * undefined Remote host closed the connection
1138003974 J * undefined ~undefined@adsl-68-93-109-94.dsl.rcsntx.swbell.net
1138004381 Q * shedi Quit: Leaving
1138006556 Q * Pazzo Remote host closed the connection
1138006620 J * balbir ~balbir@59.145.136.1
1138006686 P * undefined 
1138007666 J * prae ~prae@ezoffice.mandriva.com
1138007688 J * meandtheshell ~markus@85-124-35-76.dynamic.xdsl-line.inode.at
1138007788 J * shedi ~siggi@tolvudeild-205.lhi.is
1138007847 M * lonewolff morning all
1138007859 Q * Aiken__ Ping timeout: 480 seconds
1138008573 J * yang ~boni@cpe-213-157-253-172.dynamic.amis.net
1138008872 J * Pazzo ~Pazzo@host130-250.pool8172.interbusiness.it
1138011341 M * lonewolff is there anyone about who knows much about disk limits? im having some probelms :(
1138011444 M * daniel_hozac shoot.
1138011526 J * tAz LsJz@81.169.234.158
1138011543 A * tAz s e r v e r _ i r c . g a r i p n e t . c o m
1138011545 P * tAz 
1138011605 M * lonewolff well i have remouted my vserver partition with the tagxid option, created a new vserver and set up some limits are per the documentation on disk limits, but im not sure how to find out if its worked (df -h still shows full drive), so i tried using chxid -c <vservername> -R /var/lib/vservers<vsername> and then teh guest refuses to start
1138011634 M * lonewolff basically looking for a bit of guidance as to what should / shouldnot happen etc
1138011687 M * daniel_hozac you have a static context for the guest, right?
1138011712 M * daniel_hozac df -h inside the vserver should show the limited amount of space and inodes.
1138011745 M * lonewolff i think i see what my problem is, what is a static context?
1138011790 M * daniel_hozac a number that identifies the guest, that is persistent across starts/stops of the guest.
1138011805 M * daniel_hozac configured in /etc/vservers/<name>/context.
1138011808 M * lonewolff it is mentioned on the howto im following, is it just a case of putting a number in /etc/vservers/<name>/context or is there more to it than that
1138011845 M * daniel_hozac nope, that's it,
1138011927 M * daniel_hozac are you following http://linux-vserver.org/Disk+Limits or something else?
1138011936 M * lonewolff yes following that
1138011994 M * daniel_hozac you have util-vserver 0.30.210 too, right?
1138012112 M * lonewolff ah i seem to have 0.30.209-2 (from debain)
1138012194 M * daniel_hozac disk limits aren't supported in the configuration until 0.30.210.
1138012264 M * lonewolff ok, ill have to compile tha tthen
1138012564 M * lonewolff might have more luck when actually suing a version of the prog which supports what im doing lol
1138013776 M * daniel_hozac heh.
1138014470 J * azazel ~azazel@81-174-46-248.f5.ngi.it
1138014620 J * DavidS ~david@chello062178045213.16.11.tuwien.teleweb.at
1138014676 M * DavidS hi, after being quite frustrated with the wiki and some fatherly advice from Bertl_zZ, I have begun refactoring it (the wiki)
1138014708 M * DavidS I started with the documentation and help sections on the main page and the short presentation. Feedback welcome
1138015467 M * DavidS are the 2.4 patches still developed or only in "forward-porting-mode"?
1138015483 M * daniel_hozac they are in maintenance mode.
1138016025 M * DavidS daniel_hozac: so, no new features but only security fixes and "rediffing" against newest 2.4 release
1138016045 M * daniel_hozac DavidS: pretty much.
1138016051 M * DavidS thanks
1138018522 M * DavidS now i did the downloads thingy too ... some links to 2004 stuff, i just killed, others i have moved around a bit. 
1138018552 M * DavidS the debian stuff for example is outdated and redundant, I assume the Redhat pieces aren't much better.
1138018585 M * DavidS would someone sponsor a common repository with good (enough) connectivity to put something mor structured up?
1138018644 M * azazel how much connectivity?
1138018710 M * DavidS azazel: hmm, "depends". compressed minimal guest images have ~ 40-70Megs.
1138018735 M * DavidS but perhaps it is just needed to throw away the old stuff .. i'm a wimp in this regard ;)
1138018759 M * DavidS azazel: it's probably more a question of maintenance than bandwidth anyways ...
1138019149 M * daniel_hozac what Red Hat pieces?
1138019794 M * yang DavidS: you want to make a newer version of debootstrap?
1138019907 M * DavidS yang: why?
1138019921 M * yang or what were you talking about?
1138019981 M * yang you need a ftp for downloading debian debootstrap process?
1138020013 M * aba yang: debootstrap is a small shell script that uses wget to download the final parts
1138020021 M * yang yes
1138020027 M * aba of course, there are also ready root images available ...
1138020049 M * DavidS yang: no, I was only a bit appalled by the quite unstructured mass of only semi-useful images in the prebuilt section of linux-vserver.org
1138020054 M * yang but i am asking him if he need a ftp to put this images on it?
1138020072 A * aba doesn't understand that question
1138020123 M * DavidS aba: yang thinks i have made new debian guests, because i complained about those on the linux-vserver page earlier
1138020147 M * aba DavidS: ah, ok.
1138020148 M * daniel_hozac well, Debian is the least problematic distro to build.
1138020176 A * DavidS puts on his documentation writer hat
1138020197 M * DavidS the problem is not building the images, but the presentation on http://linux-vserver.org/Linux-VServer
1138020240 M * daniel_hozac DavidS: hmm, why did you put an old util-vserver version on the front page?
1138020307 M * DavidS daniel_hozac: because that .209 is what is linked from http://www.13thfloor.at/vserver/s_rel26/v2.01/
1138020348 M * daniel_hozac the release pages are typically not updated post-release.
1138020370 A * DavidS is no prophet
1138020418 M * daniel_hozac util-vserver downloads are best to get from http://www.13thfloor.at/~ensc/util-vserver/files/alpha/
1138020551 M * DavidS done *sigh*
1138020803 M * DavidS are there any plans to forward port the stable releases to current kernels (i.e. 2.6.15)?
1138020829 M * daniel_hozac what?
1138020833 M * daniel_hozac of util-vserver?
1138020907 M * DavidS no, kernel patches ... d'oh: http://vserver.13thfloor.at/Experimental/ has 2.6.15 patches for 2.0.1.2 too ...
1138020922 M * DavidS (which does answer my question with yes)
1138020967 M * DavidS rather with "no, already done"
1138020995 A * DavidS feels stupid.
1138021264 M * lonewolff daniel_hozac: thanks for thehelp i not have disk limits working, time to move onto quota support inside the guests :)
1138021290 M * daniel_hozac lonewolff: well... those are pretty much mutually exclusive.
1138021332 M * daniel_hozac lonewolff: because quotas inside the guests require separate volumes/partitions for each guest, and disk limits don't make much sense then.
1138021337 M * lonewolff daniel_hozac: yeah i know, im going by order of importance, more important that i can limit the space a client server can use than they can limit the space their users get ;)
1138021457 M * lonewolff ah, loks like i nee dto complile the kernel again :(
1138021987 J * Smutje ~Smutje@xdsl-87-78-42-175.netcologne.de
1138022094 Q * Smutje_ Ping timeout: 480 seconds
1138022213 M * DavidS Bertl_zZ: It'd be great if http://vserver.13thfloor.at/Experimental/ would sort by file name and not by some kind of timestamp ...
1138022244 M * daniel_hozac http://vserver.13thfloor.at/Experimental/?C=N
1138022386 A * DavidS hands daniel_hozac a large trout.
1138022493 M * DavidS anyways, i tried again to reorganize the sources section, now it has tables component/kernel version with links
1138022532 Q * balbir Quit: Leaving
1138024116 M * lonewolff right, i just compiled my kernel with virtual root device support, and nade the device /dev/vroot/0 but im following the standard non shared quota howto and am a bit lost at the vrsetup stage
1138024158 M * daniel_hozac what's the problem?
1138024220 M * lonewolff well it says use vrsetup to tell the kernel what block device you want to handle quota for, im just a little confused because all my vservers are one the same partition
1138024237 M * lonewolff im probably missing something again or my brain is just not working properly
1138024250 M * daniel_hozac as i said earlier, that won't work with quotas.
1138024266 M * daniel_hozac each guest will need a separate volume/partition.
1138024284 M * lonewolff you said that earlier :( i didntnotice that
1138024404 M * lonewolff if i put vservers on lvm volumes that wouhat would take care of the disk limit side of things without disk limits .. yes? and then i could set up quota support inside the guests?
1138024429 M * daniel_hozac exactly.
1138024434 M * lonewolff altho i am also thinking what happens if i have more than 8 guests?
1138024452 M * daniel_hozac IIRC you can raise the amount of vroots on the kernel command line.
1138024465 M * azazel more than 8?
1138024477 M * lonewolff ah cool
1138024483 M * lonewolff this has certainly been a day of learning
1138024490 M * azazel :)
1138024581 M * daniel_hozac you can raise the default in the kernel too, if you want.
1138024630 M * lonewolff right, time to hack up my large lvm /vserversinto some smaller ones
1138024763 Q * pusling Remote host closed the connection
1138024775 J * pusling pusling@195.215.29.124
1138025875 J * Viper0482 ~Viper0482@p549760D5.dip.t-dialin.net
1138026343 J * mkhl ~mkhl@200-153-153-53.dsl.telesp.net.br
1138031625 N * Bertl_zZ Bertl
1138031650 M * Bertl morning folks!
1138031707 M * Bertl DavidS: the time stamp is default on purpose, but as usual, by selecting the filename tab, you can sort by that too .. why is the timestamp default? because it shows the _latest_ patch :)
1138032034 M * daniel_hozac morning!
1138032061 M * Bertl DavidS: ad reorganization, nice, but somewhat misleading ...
1138032065 J * Milf ~Miranda@ipsio306.ipsi.fraunhofer.de
1138032110 M * Bertl DavidS: also, I hope you plan to update that everytime I release something, otherwise it will be outdated in a few hours ...
1138032686 J * zobel zobel@2001:7b8:385::2
1138032689 M * zobel hi
1138032731 M * zobel on debian i run into the following problem when building a chroot using vserver command:
1138032783 M * zobel vserver $vservername build -m debootstrap -- -d sarge -m http://amd64.debian.net/debian-amd64/debian -- --resolve-deps
1138032801 M * zobel line 206: -n: command not found
1138032810 M * daniel_hozac zobel: update to 0.30.210.
1138032814 M * Bertl what version do you use?
1138032853 M * zobel 0.30.209-0bpo1 (backports.org)
1138032865 M * Bertl okay, pls give 0.30.210 a try
1138032872 M * zobel okay, will try the 0.30.210
1138033106 Q * Milf Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org
1138033160 M * bubulak Bertl: its passible to setup vserver on virtual interface like eth0:1 or dummy0:0 ? 
1138033166 M * bubulak possible
1138033202 M * Bertl sure, you just have to specify 'name' in the interfaces/*/
1138033220 M * bubulak hmm ok i try
1138033228 M * Bertl or use the <name>=eth0.... syntax on --interfaces
1138033244 M * Bertl the name in your case is 'just' the '0' 
1138033261 M * Bertl so a name of hansi would give you eth0:hansi
1138033321 M * bubulak work :)
1138033328 M * bubulak thankz
1138033343 M * bubulak i tryed other way
1138033351 M * bubulak echo dummy0:0 > dev
1138033356 M * bubulak thats don't work
1138033450 M * Bertl well, I did not say that this would work, did I?
1138033469 M * Bertl once again: if you want a name like dummy0:0 then
1138033481 M * Bertl you set 'dev' to dummy0 (which is the device)
1138033495 M * Bertl and 'name' to '0' which is the alias for this device
1138033519 M * bubulak ok 
1138033525 M * bubulak thankz
1138033529 M * Bertl np
1138033658 M * Bertl okay, off for a while .. back later ...
1138033666 N * Bertl Bertl_oO
1138033862 J * stefani ~stefani@superquan.apl.washington.edu
1138034111 P * stefani 
1138034779 M * zobel yay, 0.30.210 doesn't build properly :/
1138034840 M * zobel make[1]: Entering directory `/usr/src/util-vserver-0.30.210/man'
1138034840 M * zobel make[1]: *** No targets specified and no makefile found.  Stop.
1138034840 M * zobel make[1]: Leaving directory `/usr/src/util-vserver-0.30.210/man'
1138034840 M * zobel make: *** [build-stamp] Error 2
1138034954 Q * DavidS Quit: Download Gaim: http://gaim.sourceforge.net/
1138034990 Q * bubulak Quit: leaving
1138035193 M * daniel_hozac zobel: builds fine here.
1138035431 M * Loki|muh i have a stramge problem with apache and vserver
1138035450 M * Loki|muh requests are interrupted after byte 11711
1138035453 M * Loki|muh see here: http://pastebin.com/519249
1138035506 M * Loki|muh anyone known something about that?
1138035703 Q * shedi Quit: Leaving
1138036036 J * bonbons ~bonbons@83.222.39.249
1138036504 Q * prae Quit: Execute Order 69 !
1138037262 M * daniel_hozac Loki|muh: 2.1.0*, right? disable sendfile.
1138037351 M * Hollow scnr: http://dev.croup.de/vwiki/index.php/ ;)
1138037383 M * Loki|muh daniel_hozac: http://list.linux-vserver.org/archive/vserver/msg11831.html <-- this problem here?
1138037426 M * daniel_hozac Loki|muh: i think that should already be included in the most recent patches.
1138037453 M * Loki|muh okay, do you know which patch is recommended?
1138037488 M * daniel_hozac the latest of whichever series you follow ;)
1138037510 M * Loki|muh okay :)
1138037569 M * daniel_hozac Hollow: nice.
1138037588 M * Hollow thanks :)
1138037661 M * Loki|muh is there a default ulimit inside vservers?
1138037693 M * FaUl the same as on the host 
1138037822 M * Loki|muh thanks
1138037963 Q * monrad Ping timeout: 480 seconds
1138038820 J * monrad ~mikkel@213083190131.sonofon.dk
1138039073 M * Loki|muh daniel_hozac: do you know if patch-2.6.15-vs2.1.0.5.diff ist compatibel with 2.6.15.1?
1138039832 J * prae ~benjamin@sherpadown.net
1138039899 J * click_ click@ti511110a080-5108.bb.online.no
1138039899 Q * click Read error: Connection reset by peer
1138039963 M * Loki|muh seems so
1138040000 M * Loki|muh is VSERVER_IDLETIME experimental and unstable? or are there other reasons why it is disabled per default?
1138040353 Q * pusling Read error: Connection reset by peer
1138040356 J * pusling_ pusling@195.215.29.124
1138041970 J * MakMoer ~makmoer@213.225.59.162
1138042040 M * MakMoer HI all..I am trying to get openvpn running in a vserver..does anyone have experience with this? I am getting the error : SIOCSIFFLAGS: Cannot assign requested address
1138042081 M * daniel_hozac MakMoer: CAP_NET_ADMIN? you haven't limited the guest to any IPs, have you?
1138042106 M * MakMoer CAP_NET_ADMIN is in my bcapabilities file..
1138042113 M * MakMoer the guest has one IP..
1138042190 M * MakMoer in bcapabilities I have CAP_NET_ADMIN, CAP_NET_RAW and CAP_MKNOD for this host..I am a bit clueless as to what else I need to do..
1138042467 N * Bertl_oO Bertl
1138042505 M * Bertl MakMoer: basically you don't need CAP_MKNOD, but you want to assign 0.0.0.0 as one and only IP to the guest
1138042524 M * Bertl MakMoer: ah, and you should have a very recent patch
1138042571 M * Bertl Loki|muh: well, it is development ... and it is part of the new fair scheduler so not that well tested (yet)
1138042577 M * MakMoer ok..at the moment this guest is 192.168.1.25, do I need to change this to 0.0.0.0 ? which patch..
1138042621 M * Bertl Loki|muh: but, as scheduler issues usually let your system explode within a few minutes, I'd consider it somewhat safe ...
1138042650 M * Bertl MakMoer: IIRC, the 0.0.0.0 binding was introduced in one of the recent devel patches
1138042705 M * Bertl daniel_hozac: what's the problem with sendfile?
1138042758 M * Bertl zobel: but 0.30.209 did?
1138042808 M * MakMoer Bertl : Sorry..I am afraid I am still clueless..
1138042810 M * daniel_hozac Bertl: i assumed it's the same issue Doener patched.
1138042883 M * daniel_hozac (a bit premature, i guess)
1138042892 M * daniel_hozac easy to test if it is the case though.
1138042952 M * Bertl MakMoer: what kernel patch do you use?
1138042975 M * Bertl daniel_hozac: because Doener's patch should be in recent devel
1138042990 M * daniel_hozac Bertl: yeah, that's what i figured.
1138043159 M * MakMoer Bertl : I followed the howto : http://linux-vserver.org/VServer+installation+Fedora+Core+4 and used "Installing the host kernel from rpm"
1138043170 M * zobel Bertl: i found the problem with 0.30.209 for debian-backports.
1138043191 M * Bertl MakMoer: okay, which rpm is mentioned there?
1138043197 M * MakMoer Bertl : Daniel Hokka Zakrisson's packages
1138043198 M * zobel if /etc/vservers/$vservername already exists, the script aborts...
1138043282 M * Bertl daniel_hozac: what would MakMoer get when following the how-to?
1138043303 M * daniel_hozac Bertl: 2.6.14 based with 2.0.1 + PAT-2.0.2.
1138043325 M * MakMoer Bertl : Linux vserv 2.6.14-1.1656_FC4.vs2.0.1.2
1138043522 M * MakMoer Bertl : I also found this thread http://list.linux-vserver.org/archive/vserver/msg09858.html that states I need : ~hide_netif in 'flags'
1138043664 M * MakMoer Bertl : ahh...that seems to make it work..
1138043687 M * MakMoer on the host I now see : tun0      Link encap:UNSPEC inet addr:10.8.0.1  P-t-P:10.8.0.2  Mask:255.255.255.255
1138043694 P * zgrim left
1138043736 M * Bertl MakMoer: aha, interesting ... IMHO the kernel you have should not allow binding to arbitrary IPs ...
1138043782 M * MakMoer Bertl : hmm..interesting..I will see if it all works..
1138043805 M * Bertl nevertheless, you can always specify the tun IP if you know it
1138044728 J * Borg_Number_One ~BorgNr1@p54B8A574.dip0.t-ipconnect.de
1138044736 M * Borg_Number_One Good evening. :)
1138044991 Q * jgommers Ping timeout: 480 seconds
1138045355 M * Bertl good evening B#1!
1138046192 M * Loki|muh Bertl: thx
1138046300 M * Bertl Loki|muh: np, btw, you need an experimental tool to make use of the idle bucket extension
1138046575 J * liquid3649_ ~Viper0482@p54975F39.dip.t-dialin.net
1138046650 M * Hollow Bertl: http://dev.croup.de/vwiki/
1138046664 Q * prae Quit: Pwet
1138047009 Q * Viper0482 Ping timeout: 480 seconds
1138047068 M * Bertl welcome liquid3649_!
1138047110 M * Bertl Hollow: cool, want some comments?
1138047114 M * Hollow sure
1138047141 M * Bertl first, we should get rid of the excessive long pathes
1138047152 M * Bertl (as we did in tavi)
1138047161 M * Hollow yeah.. just too lazy for mod-rewrite for testing mediawiki ;)
1138047188 M * Bertl np, you can have my rewrite rules for tavi, as I guess it requires some mediawiki support too
1138047196 J * undefined ~undefined@adsl-68-93-109-94.dsl.rcsntx.swbell.net
1138047204 M * Bertl (otherwise mediawiki will use long links)
1138047205 M * Hollow afaik there are some rules shipped with mediawiki..
1138047220 M * Hollow just didn't use it
1138047227 M * Bertl we should definitely test/fix this before we move anything there
1138047256 M * Bertl then, would it be possible to make the 'news' section tighter?
1138047277 M * Bertl (similar to the news we have now)
1138047306 M * Hollow sure, i did it with templates so it should be quite easy
1138047326 M * Bertl the navigation area and toolbox seems quite fine to me
1138047351 M * Hollow i basically put the news + version box and this little text + logo on it, just to see what it looks like
1138047355 M * Bertl also the recent releases and the 'Introduction' (what is this entry called btw?)
1138047358 M * Hollow but i think mediawiki is quite nice
1138047397 M * Bertl and, could we get the footer a little tighter too?
1138047420 M * Hollow sure, will update it and ping you again
1138047421 M * Bertl i.e. just as high as the GNU_FDL/Media Wiki Icon?
1138047440 M * Hollow k
1138047476 M * Bertl line spacing seems a little too large too (for the normal text and bullet lists
1138047511 M * Bertl btw, what do you think of the wireframe penguin idea?
1138047568 M * Bertl ah, two other things:
1138047604 M * Bertl IMHO it would look better if the 'Create account' would be at the same line as the article, discussion, edit ...
1138047631 Q * _Roey Remote host closed the connection
1138047657 M * Bertl and I would really like to have Linux-VServer somewhere on the right (in the header), e.g. right of the title or so ...
1138047774 M * azazel Bertl: and others, what about a trac instance with integrated bug tracker and all the other stuff?
1138047790 M * Bertl would also be worth a try to use a font with serifs ...
1138047808 M * Bertl azazel: I think we can add a lot of stuff, but we better do it step by step
1138047830 M * Bertl otherwise we end up with a bunch of features nobody uses
1138047850 M * Bertl sidenote: we even have a bugzilla somewhere, but it is not used by anyone :)
1138047889 M * azazel Bertl: trac is first an useful wiki...
1138047919 M * Bertl well, advantages over mediawiki?
1138047921 M * azazel as a sidenote (do you found the slides?:))
1138047938 M * Bertl to be honest, I didn't find the time to search for them
1138048181 M * azazel nothing really that belong to the wiki side. but it integrates an svn/darcs viewer, a bug tracker , and a timeline, to see changes and roadmaps. The useful thing is that all is referenceable by wikilinks, svn revisions, changes, bugs, milestones and so on
1138048239 M * azazel is/wikilinks,/wikilinks:
1138048239 M * Bertl well, we don't use svn/darc, we do not use a bug tracker yet (except for savannah/utils) and I hate timelines :)
1138048260 M * Bertl but I appreciate the suggestion ...
1138048282 J * shedi ~siggi@inferno.lhi.is
1138048356 J * Aiken ~james@tooax8-147.dialup.optusnet.com.au
1138048360 M * azazel also, with trac it's possible to use reStrucutured text syntax, which is very well formed and interesting (http://docutils.sf.net)
1138048415 M * ebiederm Bertl:  I just threw a patch out for discussion, now to see where it goes...
1138048425 M * Bertl url?
1138048464 M * ebiederm That would have been better.  I just sent it out via emal.
1138048496 M * ebiederm I copied the vserver list though...
1138048502 M * Bertl ah, good!
1138048548 M * ebiederm Anyway I successfully implemented pids as weak references in the kernel....  Which solves one of the classic pid problems.
1138048569 M * Bertl well done, but without virtualization yet, right?
1138048573 M * ebiederm And the work is very similiar to what it takes to virtualize pids :)
1138048605 M * ebiederm Bertl:  Yes without virtualization yet.  Although that patch becomes much smaller afterwards.
1138048614 M * Bertl of course ...
1138048698 M * Bertl my bme patch-set gets smaller and smaller with every kernel release ... there are many trivial patches from Al Viro and recently HCH ....
1138048777 M * ebiederm I am having the hardest time imagining how to trim down patches that have such a global effect.
1138048782 M * Bertl ebiederm: is this your first mail to the VSML?
1138048812 M * ebiederm Bertl:  It should be my second.  My first you refered me to irc!
1138048828 M * Bertl thought so, just because it didn't arrive there yet
1138048856 M * ebiederm Well I think my email was a little on the large side...
1138048872 Q * mkhl Ping timeout: 480 seconds
1138048940 M * ebiederm Anyway I'm going to take a break for a bit.  
1138048969 M * Bertl okay, any plans for the networking stuff yet?
1138049024 M * ebiederm Bertl:  Getting there.
1138049039 M * Bertl okay, end of the week or so?
1138049047 M * ebiederm Something like that.
1138049060 M * Bertl good, just that I can reserve some time ...
1138049060 M * ebiederm The networking stuff should actually be a lot simpler as I can do full separation.
1138049100 M * ebiederm Needing to support waitpid accross pid spaces is a lot less straight forward.
1138049143 M * Bertl I can see that ...
1138049164 M * ebiederm With pidspaces logic needs to be changes where I only need to changes which values I am passing in the networking code.
1138049229 M * ebiederm There are a lot of changes on the networking side still to worry about.
1138049273 M * ebiederm I keep staring at my typing, seeing the typos and not believing I'm a native English speaker!
1138049352 M * Bertl you are? :)
1138049421 M * Bertl btw, just found HTML on your resume page (from 2000) under 'Comfortable Programming in:' .. so could you point me to a program of yours in HTML?
1138049454 M * ebiederm No comment.
1138049503 M * Aiken Bert starting to think the kernel from yesterday is not stable
1138049514 M * Bertl ah, in what way?
1138049538 M * Aiken it's 2nd reboot with no oops and no log messages within the last 10 minutes
1138049568 M * Bertl hmm, okay, that doesn't sound very stable indeed ...
1138049581 M * Bertl but I guess we have to expect something like that from a git version
1138049597 M * Bertl let's wait for rc2 with further testing, no?
1138049598 M * Aiken at least the git version compiles unlike their rc
1138049745 M * ebiederm Well at least report the bug.
1138049752 M * MakMoer Bertl : My openvpn is working fine..but I have a openvpn routing question..Not sure it is relevant to this group..can I ask ?
1138049765 M * Bertl sure, go ahead!
1138049825 M * Bertl ebiederm: guess 'my git kernel reboots on my alpha' is not a good bug-report, and I'm pretty sure Aiken will report anything more substantial ...
1138049844 M * Aiken I'll report something when I know what
1138049861 M * Aiken a report of 'it reboots for no reason' is useless
1138049931 M * MakMoer ok..my vserver has the IP : 192.168.1.25 and the openvpn server IP is : 10.8.0.1. I can connect to the vpn from home and I get the IP in windows : 10.8.0.5. Now.My vserver which is the openvpn server can ping another server on the office lan. 192.168.1.1 but I cannot access it from home when connected to the vpn. I can however ssh to other vservers on the host where the openvpn vserver is running..
1138050009 M * MakMoer Bertl : I think I need to do this : Next, you must set up a route on the server-side LAN gateway to route the VPN client subnet (10.8.0.0/24) to the OpenVPN server (this is only necessary if the OpenVPN server and the LAN gateway are different machines).....but no idea how/where
1138050081 M * Bertl well, typically I'd say you want to use masquerading here
1138050088 M * Bertl something like:
1138050122 M * Bertl iptables -t nat -I POSTROUTING -s 10.8.0.0/24 -d 192.168.1.0/24 -j MASQUERADE
1138050152 M * MakMoer this is now on the vserver running openvpn ?
1138050154 M * Bertl that will work for all IPs in the 10.8.0.0/24 range except for the guest IP
1138050165 M * Bertl on the host
1138050172 M * MakMoer ok
1138050224 M * Bertl the client on the other hand has to use the 10.8.0.1 ip as gateway
1138050305 M * MakMoer ok.thanks..
1138050388 M * Bertl you're welcome!
1138050546 M * MakMoer ok..and I needed : echo 1 > /proc/sys/net/ipv4/ip_forward
1138050586 M * MakMoer working perfectly now..great..
1138050735 M * MakMoer damn..you guys are the best man !!
1138050819 J * Roey ~katz@h-69-3-4-130.mclnva23.covad.net
1138050832 M * Bertl welcome Roey!
1138050842 M * Roey welcome Bertl!! :)
1138050842 M * Bertl MakMoer: tx for the flowers!
1138050922 M * MakMoer Bertl : My pleasure..it is great talking to people who know what they're doing..
1138050983 M * Aiken finally got an oops http://pastebin.com/519682
1138051026 M * Bertl well, that looks very similar to what we had yesterday, except for a new location
1138051070 M * Roey anyone here do OpenVPN?
1138051079 M * Aiken pc arch/alpha/mm/fault.c:91     ra arch/alpha/kernel/entry.S:140
1138051083 M * Bertl Roey: check with MakMoer!
1138051144 M * MakMoer Roey>: Yes..I just got it working 10 mins ago..he he
1138051174 M * Roey heh
1138051176 M * Roey MakMoer:  hi!
1138051180 M * Bertl Aiken: hmm, no, that one is different ... seems that struct mm got lost!
1138051187 M * Roey MakMoer:  listen.. why the hell does OpenVPN only start ONE service, when I have TWO conf files?
1138051189 M * Roey :)
1138051206 M * Aiken all I was only doing a flood ping
1138051263 M * MakMoer Roey: Sorry, I am not an advanced openvpn user..I just got it working with one config..
1138051280 M * Aiken the other error is   eth0 DMAing conflict in net2k_pci_ (something mabye _get_8390 and block_?)
1138051287 M * Bertl daniel_hozac: ping?
1138051406 M * Roey oK I found it out.
1138051409 M * Roey MakMoer:  I figured it out
1138051412 M * Roey /etc/default/openvpn
1138051419 M * Roey only had two things listed under AUTOSTART
1138051632 M * MakMoer ok..I dont even have a /etc/default/openvpn
1138051652 M * MakMoer but mine is manual at the moment..
1138051917 J * vrwttnmtu ~eryktyktu@82-69-161-137.dsl.in-addr.zen.co.uk
1138052030 M * Bertl welcome vrwttnmtu!
1138052156 M * vrwttnmtu Hey Bertl 
1138052161 M * vrwttnmtu Ca va?
1138052165 M * vrwttnmtu Or Wie Gehts?
1138052171 M * vrwttnmtu Kak dela?
1138052179 M * vrwttnmtu Hmm
1138052182 J * mkhl ~mkhl@200-148-41-194.dsl.telesp.net.br
1138052183 M * vrwttnmtu How many others?
1138052187 M * vrwttnmtu Como estas?
1138052192 M * vrwttnmtu I think I'm out now
1138052199 M * Bertl I'm fine, and I guess we all know english good enough :)
1138052210 M * Bertl welcome mkhl!
1138052213 M * vrwttnmtu Lucky for me :)
1138052227 M * vrwttnmtu mkhl, What's "How are you in Brazilian"?
1138052237 M * vrwttnmtu I mean "How are you?" in Brazilian
1138052291 M * vrwttnmtu So Bertl, what's new in vserver world?
1138052407 M * Bertl 2.6.16-rc1 
1138052546 Q * MakMoer Quit: 
1138052621 M * vrwttnmtu No, no :) I meant features, improvements, etc :)
1138052704 M * Bertl ah, well, we have an improved scheduler, page fault and slab accounting, a few cleanups and soft limits ...
1138052716 M * Bertl (soft limits are in an early stage though)
1138052737 M * vrwttnmtu soft limits.. :)
1138052750 M * vrwttnmtu Excellent :)
1138052785 M * vrwttnmtu Don't know if the scheduler pages and slabs make much difference to me
1138052786 M * vrwttnmtu :)
1138052958 M * Bertl well, not everything makes a difference to everybody ...
1138053212 M * Roey Bertl:  I'm soooo waiting for openvpn support in vserver.
1138053234 M * Bertl talk to ebiederm ... :)
1138053276 M * Roey ebiederm:  hi
1138053296 M * Roey ebiederm:  are you this guy
1138053298 M * Roey  making the possibility for runni
1138053303 M * Roey http://linux-vserver.org/Todo+List
1138053307 M * Roey the Keeper of that list?
1138053310 M * Bertl (he is laying the foundation for ngnet, hopefully)
1138053316 M * Roey Bertl:  hey, you're going for commercialization of openvpn? 
1138053328 M * Roey Bertl:  there could be a lot of potential for management
1138053332 P * meandtheshell 
1138053334 M * Roey Bertl:  what's ngnet?
1138053341 M * Roey Bertl:  next-gen networking for the next vserver?
1138053344 M * daniel_hozac Bertl: pong
1138053364 Q * wibble Ping timeout: 480 seconds
1138053387 M * Bertl daniel_hozac: I have a problem with a strange fc2 installation, I'd like to install an fc4 guest there, but it seems to fail regardless what I try
1138053419 M * Bertl first, there seems to be no patch for the fc2 yum
1138053429 M * ebiederm Roey:  I'm working in that general vein.
1138053440 M * daniel_hozac FC2 is yum 2.1, isn't it?
1138053460 M * Bertl rpm -qa | grep yum
1138053460 M * Bertl yum-2.0.7-1.1
1138053467 M * daniel_hozac heh, even worse.
1138053479 M * Bertl but I guess 2.1 something would work
1138053493 M * daniel_hozac yum isn't much for backwards compatibility, and the repository format changed from 2.1 to 2.2 (IIRC).
1138053494 M * Bertl but recompiling does not work, as it complains about a missing gettext
1138053516 M * daniel_hozac 2.2 being used in FC3+.
1138053528 M * Bertl so how to proceed? 
1138053537 Q * vrwttnmtu Ping timeout: 480 seconds
1138053544 M * Bertl I've got apt-rpm too, but that fails too
1138053560 M * ebiederm Roey:  What is openvpn?
1138053569 M * Bertl daniel_hozac: btw, yum seems very inefficient to me ...
1138053591 M * Bertl daniel_hozac: it takes about 3 minutes to figure that it cannot install a few packages)
1138053596 M * daniel_hozac Bertl: that's one of the changes in 2.2+.
1138053620 M * Roey ebiederm:  it's an SSL vpn
1138053635 M * daniel_hozac are there any apt repositories around for FC4?
1138053646 P * Borg_Number_One 
1138053651 M * daniel_hozac i think apt-rpm is discontinued upstream as well.
1138053669 M * Roey ah
1138053694 M * Bertl daniel_hozac: ah, so they are finally using urpmi :)
1138053720 M * ebiederm Roey:  What kernel services does it depend on?
1138053732 M * Bertl ebiederm: tun/tap devices
1138053763 M * Roey ebiederm:  hermmm it need to be able to open and close tun/tap devices
1138053768 M * Roey ebiederm:  as well as add and remove routes
1138053779 M * daniel_hozac Bertl: hah, no. skvidal would never think about using something already existing ;)
1138053859 M * Bertl i.c.
1138053861 M * daniel_hozac yum also doesn't handle installing multiple versions at all, so i guess the situation is even worse than that of Debian.
1138053872 M * ebiederm Ok.  tun/tap is a bit of a pain but no real problems, adding and removing routes I have already done.
1138053933 M * ebiederm The challenging will be getting all of the pieces integrated.
1138053948 M * Roey ok
1138053963 J * vrwttnmtu ~eryktyktu@82-69-161-137.dsl.in-addr.zen.co.uk
1138053982 M * ebiederm err integrated with vserver.  
1138053997 M * Roey ebiederm:  btw keep up the good work; I like this project more and more
1138054008 Q * liquid3649_ Remote host closed the connection
1138054009 M * Roey who cares about overbloated Xen which doesn't really increase security
1138054020 M * Roey (and is not a direct competitor anyhow)
1138054101 Q * bonbons Quit: Leaving
1138054300 N * ebiederm ebiederm_oO
1138054370 J * wibble wibble@vortex.ukshells.co.uk
1138054378 M * Bertl welcome wibble!
1138054562 Q * Roey Quit: Leaving
1138054789 Q * ntrs_ Remote host closed the connection
1138054887 M * Aiken trying an attribute statement in mm_struct
1138054910 M * Loki|muh is there a trick with rlimits like /etc/vservers/pbuilder1/rlimits/nofile? 
1138054929 M * Loki|muh i did a echo 4096 > /etc/vservers/pbuilder1/rlimits/nofile and restarted the vserver but nothing changed
1138054954 M * Loki|muh uh, i should change /etc/security/limits.conf inside the vserver, too, right?
1138055012 J * wibble_ wibble@vortex.ukshells.co.uk
1138055102 Q * wibble Ping timeout: 480 seconds
1138055219 Q * vrwttnmtu Remote host closed the connection
1138055221 M * Loki|muh hmmm no, no effect :(
1138055789 M * Bertl what do you expect to change?
1138055815 M * Bertl I mean, what does /proc/virtual/<xid>/limit show for your guest?
1138056024 M * Loki|muh hmmm FILES:       210               0/     231           4096/    4096            0
1138056089 M * Bertl so, you are using 210 of 4096 maximum
1138056100 M * Bertl and the highest observed value is 231
1138056114 M * Loki|muh i expect that ulimit -n gives me 4096 as return
1138056129 M * Bertl why?
1138056150 M * Bertl the ulimit is something different, set the ulimit resources for that
1138056170 M * Bertl ulimit is 'per user'
1138056240 Q * mnemoc Read error: Connection reset by peer
1138056248 M * Cru morning
1138056248 M * Loki|muh hmm, so that is independent from vserver configuration?
1138056313 M * Bertl well, there is an entry called ulimits (similar to rlimits) in the config
1138056321 M * Bertl which sets the 'defaults'
1138056331 J * mnemoc ~amery@200.75.27.76
1138056349 M * Bertl morning Cru!
1138056430 M * Bertl daniel_hozac: where do I get the fc4 keys from?
1138056445 M * Loki|muh Bertl: the flower page says, that this is 2.4 only
1138056553 M * Bertl well, I filed a bug report/feature request long ago, and IIRC enrico included it for recent versions
1138056671 M * Loki|muh oh, good to know 
1138056675 M * Loki|muh i will try
1138057786 M * Bertl daniel_hozac: nevermind I disabled the check
1138058814 Q * FireEgl Ping timeout: 480 seconds
1138058974 M * Loki|muh Bertl: yes, /etc/vservers/vserver-name/ulimits is working for 2.6 kernel, too. Thanks for the hint :)
1138059362 M * Bertl great!, please send an email to Enrico that he updates the flowerpage accordingly
1138059654 M * romke hmm, emerge util-vserver fails with "make[2]: *** [src/keep-ctx-alive] Error 1" --> http://romke.net/paste/ee067eb7d35978d59566df35f4d2ca8b <-- any ideas why?
1138059936 M * Bertl Hollow:?
1138060108 J * FireEgl Atlantica@Atlantica.DollarDNS.Net
1138060114 M * Bertl welcome FireEgl!
1138060692 Q * mountie Ping timeout: 480 seconds