1152663789 Q * lehkor Read error: Connection reset by peer
1152663789 J * lehkor ~lehkor@ns.sampo.ru
1152664096 Q * kaner Write error: connection closed
1152664102 J * kaner kaner@strace.org
1152665495 J * mire ~mire@156-166-222-85.COOL.ADSL.VLine.verat.net
1152666406 J * gerrit ~gerrit@67.160.146.170
1152666541 Q * yarihm Quit: Leaving
1152667013 J * coocoon ~coocoon@84.160.101.243
1152670807 Q * mountie Quit: LUNCK!
1152672166 J * lehkor_ ~lehkor@ns.sampo.ru
1152672166 Q * lehkor Read error: Connection reset by peer
1152678076 J * Viper0482 ~Viper0482@p54976B25.dip.t-dialin.net
1152681487 Q * Viper0482 Remote host closed the connection
1152682214 J * Milf ~Miranda@ipsio96.ipsi.fraunhofer.de
1152682224 M * Milf 'LO
1152682834 Q * anonc Remote host closed the connection
1152683322 J * meandtheshell ~markus@85-125-230-35.dynamic.xdsl-line.inode.at
1152683436 Q * Milf Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org
1152683739 Q * meandtheshell Quit: bye bye ... 
1152683745 Q * Nam Remote host closed the connection
1152683977 J * Milf ~Miranda@ipsio96.ipsi.fraunhofer.de
1152684902 J * meandtheshell ~markus@85-124-232-65.work.xdsl-line.inode.at
1152685504 J * anonc ~anonc@203.26.95.33
1152685978 J * dna ~naucki@dialer-182-43.kielnet.net
1152686443 J * pisco ~pampel@80.135.134.149
1152686495 P * pisco 
1152687344 Q * Aiken Ping timeout: 480 seconds
1152687443 J * ||Cobra|| ~cob@146.50.22.204
1152687813 J * pisco ~pampel@80.135.134.149
1152688800 N * otaku42_away otaku42
1152688808 Q * gerrit Ping timeout: 480 seconds
1152689618 Q * pisco Ping timeout: 480 seconds
1152689709 J * Smutje_ ~Smutje@xdsl-84-44-244-214.netcologne.de
1152689824 Q * Smutje Ping timeout: 480 seconds
1152689824 N * Smutje_ Smutje
1152690581 Q * locksy Ping timeout: 480 seconds
1152690665 J * slava ~slava@195.22.238.42
1152690687 J * pisco ~pampel@80.135.134.149
1152691648 Q * pisco Ping timeout: 480 seconds
1152691773 M * slava Hello. After building a few vservers with success, now I've got the following error:
1152691795 M * slava vshelper.init: can not determine xid of vserver 'vserver_name'; returned value was ''
1152691814 M * slava can anyone explain me wat goes wrong ?
1152692089 M * daniel_hozac your guest doesn't start any service and so the guest exits before it's even done starting.
1152692129 J * balbir ~balbir@59.145.136.1
1152692305 M * slava Hmmm. So I should enable any service for that, right ? It's not obvious.
1152692313 M * slava I'll try
1152692430 M * slava vserver vserver_name chkconfig --list
1152692441 M * slava 'vserver ... suexec' is supported for running vservers only; aborting...
1152692635 M * daniel_hozac you could just chroot to /vservers/<name> and run chkconfig.
1152692801 M * slava Thanks alot, it works. I'm happy now -)
1152692830 M * daniel_hozac you're welcome.
1152693552 J * schimmi ~sts@port-212-202-73-176.dynamic.qsc.de
1152695743 Q * dna Quit: Verlassend
1152696570 Q * shedi Quit: Leaving
1152697167 M * anonc daniel_hozac: i hate to ask this again - what was the final decision about how to act upon utimes() with cow-linked files?
1152697173 Q * schimmi Ping timeout: 480 seconds
1152697360 J * locksy ~locksy@mrtg.sisgroup.com.au
1152697744 M * daniel_hozac anonc: i guess there hasn't really been a final decision yet, as none of the new link breaking nor IS_IMMUTABLE fixes are in.
1152697918 M * anonc daniel_hozac: cool. So assuming that we want iunlink-marked files to behave as if they were normal writeable files (ie not to be treated as immutable) then presumably one would either cow_break_link and apply the utime() to the new file, or just ignore the utime() call and return success (probably not a good idea)
1152698810 J * dna ~naucki@dialer-187-109.kielnet.net
1152699304 M * derjohn hey! On am SMP system (2*Athlon MAP, 2 GB) I have strange problems with mem allocation ond 'gfp'. Could anyone have a look at it? http://paste.linux-vserver.org/175
1152699351 M * derjohn ah and it's a  2.6.17.3-p3-vs2.1.1-rc24-64ip-squash-drbd
1152699622 J * romke ~romke@83.16.133.162
1152699905 Q * coocoon Quit: KVIrc 3.2.0 'Realia'
1152699996 J * coocoon ~coocoon@84.160.101.243
1152700026 J * schimmi ~sts@aquila.tcs.ifi.lmu.de
1152701963 M * daniel_hozac derjohn: what is that? dmesg output?
1152701975 M * derjohn yes, a dmesg output
1152702002 M * derjohn (the machine crashed all two or three days until i got aware about the dmesg 
1152702005 M * derjohn )
1152702052 M * derjohn the tyan dual athlom mobo maybe picky when i comes to kernel stuff, i just upgraded to 2.6.17.4-rc25 and set the as limits we talk about yesterday.
1152702163 M * daniel_hozac and that crashes the box?
1152702193 M * daniel_hozac looks like a regular OOM kill to me.
1152703802 M * derjohn daniel_hozac, well I only know that i had to press reset all 2-3 days. in that case there was a ever-recurring output on the console about the OOM ... I think some process tried to allocate mem and didnt stop even OOM said "no i dont have more". maybe a problem related to dual CPU?
1152703870 J * Nam ~nam@70.78.64.62
1152704660 N * Bertl_oO Bertl
1152704664 M * Bertl morning folks!
1152704669 M * FaUl morning Bertl 
1152704674 M * FaUl good to see you
1152704680 M * FaUl i've got a  feature-request ;-)
1152704687 M * Bertl ah, let's hear!
1152704752 M * FaUl it would be nice if the guest-tty/ptys would be in the host devpts as well
1152704798 M * Bertl hmm, why?
1152704805 M * FaUl because this would allow to inform every user walls (think about shutdown -h +15) or things like this
1152704819 M * FaUl every user with wall s
1152704854 M * Bertl okay, that's a good argument, what about wall inside a guest, btw?
1152704899 M * Bertl I mean, does that work as expected (i.e. msg everything inside the guest?)
1152705012 M * FaUl Bertl: i think so
1152705044 M * Bertl what about having xid=1 showing/using all the pts then?
1152705049 J * lilalinux ~plasma@dslb-084-058-233-086.pools.arcor-ip.net
1152705056 M * FaUl mhh
1152705069 M * Bertl would that help with that? i.e. what about having a special 'wall' on the host for this?
1152705082 M * FaUl then you have to write special wrapper for shutdown/etc
1152705245 M * Bertl okay, what about guest security (i.e. protection against the host?)
1152705250 M * romke morning Bertl 
1152705269 M * Bertl FaUl: and the other way round, i.e. keep host messages from the guest :)
1152705289 M * Bertl morning romke!
1152705332 M * FaUl Bertl: i think no guest should be able to talk to the host/other guests
1152705424 M * FaUl security: interesting question - you think about the case where host-user have the same uid as guest-user, right?
1152705510 M * Bertl that is one example, but basic information hiding is another, i.e. if there is a pts (from the guest) you can easily inject/capture 'secret' data from there (on the host) which isn't trivial when you have to enter the guest
1152705536 M * Bertl well, or if you are not allowed to enter it :)
1152705598 M * FaUl i think this isn't a case as guests can be manipulated from host-root in any way
1152705661 M * Bertl have you had a look at the latest lock and admin deltas?
1152705699 M * FaUl in fact if you are root on the host you may even load some kernel-module if you want - s o theere isn't real security - everything which seems to provide security is dangerous (as it creates wrong feeling)
1152705704 M * FaUl nope
1152705759 Q * sladen Ping timeout: 480 seconds
1152705955 J * sladen paul@193.28.45.41
1152706003 M * Bertl well, of course, you can do all that, it's like adding stack randomization or non executable text sections etc .. you can always work around it, but it makes it harder to get there ...
1152706697 M * Bertl FaUl: I think the 'proper' long time solution would be to have certain messages relayed to guests too, not sure that can be done in a proper way without modifying host userspace here
1152706738 M * daniel_hozac Bertl: what do you think about anonc's request?
1152706819 Q * lilalinux Ping timeout: 480 seconds
1152706819 M * FaUl Bertl: it seems to me to be the best if all vtys are in xid=1-devpts
1152706883 M * FaUl and replace shutdown with some programm which uses xid=1 for wall and add an extra vwall 
1152706884 M * Bertl daniel_hozac: you mean cow breaking on time change?
1152706901 M * daniel_hozac Bertl: right.
1152706903 M * Bertl FaUl: I agree that this is the best solution for now ...
1152706974 M * Bertl FaUl: but would be interesting if you could bring this up on lkml too, as I see a similar (or maybe even more problematic) issue with the namespace approach in the future ...
1152707023 M * FaUl i'm not so into the namsepace-approach
1152707044 M * Bertl daniel_hozac: shouldn't be too hard, although I really doubt that it is worth the efford ... (basically we could move the break to the update/change inode path)
1152707085 M * Bertl FaUl: doesn't matter, it's quite simple, you would have a 'separate' /dev/pts space for every 'guest' and another one (separate) for the host
1152707088 A * phreak`` goes looking for Bertl's qemu mini image
1152707117 M * Bertl phreak``: http://vserver.13thfloor.at/Stuff/QEMU/
1152707154 M * cattivik Bertl: hi boss:)
1152707163 M * phreak`` Bertl: those TEST_*.bz2 right ?
1152707174 M * Bertl hi cattivik!
1152707194 M * Bertl phreak``: yes, the public ones are 'working' images
1152707277 M * daniel_hozac Bertl: which path is that? and it should really be trivial with http://daniel.hozac.com/vserver/delta-cow-feat03.diff
1152707292 M * cattivik Bertl: Have you ever discussed about "hot-swap" migration for VServer guests? I mean live systems, from one hardware to another... 
1152707321 J * lilalinux ~plasma@dslb-084-058-232-251.pools.arcor-ip.net
1152707324 M * Bertl cattivik: yes, but basically it boils down to this:
1152707348 M * Bertl you usually want to do that for two reasons: load balancing and scheduled maintainance
1152707392 M * cattivik Bertl: yep!
1152707405 M * Bertl and while you can cover the maintainance part quite fine by adding a Xen layer, the load balancing case is not really realistic, as it uses _a lot_ of resources to migrate a guest
1152707450 M * Bertl we are more leaning towards a cluster solution for Linux-VServer here
1152707464 M * Bertl (which isn't that hard to do actually :)
1152707479 M * cattivik Bertl: ah, great...
1152707496 M * Bertl everybody interested in helping there can start testing right now
1152707527 M * Bertl in this case, it would be testing the ocfs2 support (as a cluster fs is a prerequisite for that)
1152707601 M * cattivik Bertl: cool...! You said "Xen layer": I know what Xen is but why do you talk of a layer?
1152707626 M * Bertl well, you basically do a dom0 with a single domU (guest) which runs a Linux-VServer kernel
1152707657 M * Bertl as Xen 3.0 allows for a high quality live migration, you can migrate the entire VServer host in case of maintainance
1152707694 M * Bertl while you have all the benefits and resource sharing for the 'real' guests
1152707817 M * cattivik Bertl: UGH! This is true magic... I hadn't think of such a solution..! :D
1152707919 M * Bertl well, we think we do not have to reinvent the wheel over and over again, we try to focus on new inventions :)
1152707994 T * Bertl http://linux-vserver.org/ | latest stable 2.01, 1.2.10, 1.2.11-rc1, devel 2.1.0, exp 2.{0.2,1.1}-rc26 | util-vserver-0.30.210 | libvserver-1.0.2 & vserver-utils-1.0.3 | He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the wiki, and we'll forget about the minute ;)
1152708013 J * vrwttnmtu ~eryktyktu@82-69-161-137.dsl.in-addr.zen.co.uk
1152708030 M * Bertl derjohn: 2.0.2: only the utsname fix, 2.1.1: switch and socket changes + fix
1152708031 M * vrwttnmtu Afternoon
1152708032 M * derjohn Bertl, gotcha! Changes from rc25 to rc26? a) stable b) devel !
1152708039 M * Bertl yes!!
1152708047 M * derjohn incredible! an answer before the question ;)
1152708055 M * Bertl welcome vrwttnmtu!
1152708064 A * vrwttnmtu waves
1152708065 M * derjohn Bertl, utsnamefix in both?
1152708081 M * daniel_hozac Bertl: no nsclean-disable?
1152708105 M * daniel_hozac Bertl: nor lock/admin for 2.1?
1152708112 M * Bertl daniel_hozac: is part of the changes with the switch
1152708126 M * Bertl the clean, not the lock/admin which requires rewrite
1152708132 M * daniel_hozac i meant for stable :)
1152708153 M * Bertl no, not yet ...
1152708171 M * daniel_hozac not yet?
1152708176 M * cattivik Bertl: wise indeed..! :]
1152708192 M * Bertl daniel_hozac: I plan to run those too through PLM now, fix up everything in a code review for stable and get that another PLM run then ...
1152708266 M * derjohn Bertl, changelog done.
1152708272 M * Bertl daniel_hozac: please double check the code yourself, if possible I like to have something ready early next week
1152708278 M * daniel_hozac vnamespace -c does still oops the kernel, right?
1152708330 M * daniel_hozac Bertl: ok, will do.
1152708332 M * Bertl I assume so .. nothing changed there but I have no problem to remove it right now, just didn't bother yet
1152708353 M * daniel_hozac ok.
1152708357 M * Bertl (as I think we will get rid of it in the code review)
1152708388 M * daniel_hozac the oops, or the function?
1152708396 M * Bertl the function :)
1152708410 M * daniel_hozac makes sense :)
1152708461 M * Bertl but maybe we should look into the oops too, might give some clues (or did you already investigate?)
1152708554 M * daniel_hozac no, i don't think i did.
1152708803 M * Bertl I love it: http://lxc.sourceforge.net/bench/
1152708898 M * doener Bertl: any idea why the 'light' results for dbench are better than the vanilla ones? Don't know about MCR, but at least for Linux-VServer that simply doesn't make sense to me
1152709012 M * Bertl no idea, but would be interesting to investigate
1152709024 M * daniel_hozac Bertl: ah, now i remember. include/linux/list.h:__list_del won't work on the first or the last mount.
1152709066 M * Bertl daniel_hozac: aha, hmm, can we hit that case with the other pathes?
1152709085 M * daniel_hozac other paths? such as?
1152709135 M * Bertl maybe I got that one wrong, where does that del happen exactly?
1152709165 M * daniel_hozac umount_unused
1152709382 M * Bertl ah, okay, so we get rid of that with the call, tx!
1152709393 M * Bertl (with removing the cleanup cmd, I mean)
1152709393 Q * balbir Quit: Leaving
1152709667 M * daniel_hozac right.
1152709870 M * cattivik Bertl: about clustering you said: "everybody interested in helping there can start testing right now": you mean starting from where?
1152710418 M * Bertl setting up a a few nodes, testing primarily ocfs2 in connection with Linux-VServer
1152710434 Q * Milf Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org
1152710517 M * vrwttnmtu daniel_hozac, When you have a sec, can you give me a hand to work out why my v6 vserver isn't v6? :)
1152710545 M * vrwttnmtu (Or anyone that can help out really)
1152710677 M * Bertl I guess for now bonbons (not here atm) is the right person ... but doener or daniel_hozac might have a clue :)
1152710723 M * vrwttnmtu Bertl, the reason that I'm asking daniel_hozac is that I've set it up before, and got it working on other servers, but on one, using his chbind6 tool doesn't seem to apply the v6 address to the vserver
1152710856 M * Bertl i.c. k, makes sense ...
1152710899 M * vrwttnmtu Bertl, Do you know why I would get the "No support for inet6" error within the vserver? (If all was compiled with Bruno's patches)
1152710931 M * Bertl that looks more like a missing patch to me (i.e. still the ipv6 blocker in the kernel)
1152710943 M * Bertl are you _sure_ you booted the patched kernel?
1152710963 M * vrwttnmtu # uname -r
1152710963 M * vrwttnmtu 2.6.16.20-vs2.1.1-rc22IPv6
1152710992 M * Bertl okay, I have to leave now, will be back later ...
1152710995 M * vrwttnmtu OK
1152711033 N * Bertl Bertl_oO
1152711263 M * daniel_hozac vrwttnmtu: what does cat /proc/self/ninfo inside the guest say?
1152711289 M * daniel_hozac vrwttnmtu: and how are you running it?
1152711300 M * vrwttnmtu 1 sec
1152711323 M * vrwttnmtu Just V4Root, and V4Bcast, and NID
1152711330 M * daniel_hozac i believe there are instances where you'd get the "No support for inet6" if you don't have an address assigned.
1152711382 M * vrwttnmtu Thing is, I've recompiled most of my boxes now with the v6 patch, and used your tool to assign the addresses, but just on this one, it doesn't work
1152711430 M * daniel_hozac vrwttnmtu: ah right, self/ninfo is the non-working one, try /proc/virtnet/*/info
1152711450 M * daniel_hozac (from the host)
1152711472 M * vrwttnmtu ID, Info, and 0:
1152711475 M * vrwttnmtu No v6
1152711476 M * vrwttnmtu :/
1152711495 M * daniel_hozac ok, how are you running chbind6?
1152711528 M * daniel_hozac (i guess you're not using the util-vserver patch?)
1152711529 M * vrwttnmtu /root/chbind6 -n 1234 -a -6 2001:xxxxxxx/64
1152711547 M * vrwttnmtu Which returns: IPv6: 2001:xxxxxxx/64
1152711559 M * vrwttnmtu And that address is assigned to eth0 on the host
1152711567 M * daniel_hozac that just means it parsed the address.
1152711583 M * vrwttnmtu Yeah, I tried it without the /64, and it barfed
1152711593 M * daniel_hozac yeah, it requires the prefix.
1152711620 M * daniel_hozac could you strace that for me?
1152711633 M * vrwttnmtu The only thing I can think is that on this kernel I haven't compiled in some option that it requires
1152711646 M * daniel_hozac you do have CONFIG_IPV6 enabled, right? :)
1152711651 M * vrwttnmtu Yeah :)
1152711652 M * daniel_hozac (or modularized)
1152711661 M * vrwttnmtu # ping6 www.kame.net -c 1
1152711661 M * vrwttnmtu PING www.kame.net(orange.kame.net) 56 data bytes
1152711661 M * vrwttnmtu 64 bytes from orange.kame.net: icmp_seq=1 ttl=49 time=275 ms
1152711661 M * vrwttnmtu --- www.kame.net ping statistics ---
1152711661 M * vrwttnmtu 1 packets transmitted, 1 received, 0% packet loss, time 0ms
1152711666 M * vrwttnmtu rtt min/avg/max/mdev = 275.980/275.980/275.980/0.000 ms
1152711684 M * vrwttnmtu I'll have to emerge strace first
1152711690 M * daniel_hozac ok.
1152711698 M * vrwttnmtu It's a rather minimal box. Give me 3
1152711756 M * vrwttnmtu daniel_hozac, join #ksdjhfkjdshf (for pasting - rather not put it on a paste site
1152711757 M * vrwttnmtu )
1152712002 N * Ben_zZz Ben_
1152712047 Q * michal` Ping timeout: 480 seconds
1152712580 J * michal` ~michal@www.rsbac.org
1152714399 J * mountie ~mountie@CPEdeaddeaddead-CM000a739acaa4.cpe.net.cable.rogers.com
1152714522 J * [PUPPETS]Gonzo gonzo@langweiligneutral.deswahnsinns.de
1152714563 Q * cattivik Quit: I'm gonna save Rose Tyler from the middle of the Dalek fleet, and then I'm gonna save the Earth
1152714615 J * TermUnitX ~PhAnATiC@dsl-201-129-207-251.prod-infinitum.com.mx
1152714737 Q * newz2000 Quit: Talk to you later
1152714835 J * yarihm ~yarihm@whitehead2.nine.ch
1152714977 M * [PUPPETS]Gonzo I have a question about the "reaper-fix" and strace: did it crash the whole host or in which way did the problem show up? And did the problem apply to gdb, too?
1152715035 M * daniel_hozac yes.
1152715047 J * newz2000 ~matt@12.216.147.124
1152715049 M * derjohn daniel_hozac, to all Questions?
1152715054 M * daniel_hozac pretty much :)
1152715070 M * derjohn daniel_hozac, does that mean every user can stop the host with gdb ?????
1152715093 M * daniel_hozac well, yes, IIRC.
1152715112 M * derjohn *ouch* 
1152715137 M * derjohn well, we should warn everyone to upgrade to newest RC to prevent DoS ?
1152715155 M * [PUPPETS]Gonzo warn noone, as people might get to know the bug ;)
1152715177 M * [PUPPETS]Gonzo Bad news that that happened, good news that I know, what caused the crashes on my production server, bad news that I searched for the reason so long ;)
1152715232 M * daniel_hozac IIRC the bug behaved differently for different people
1152715238 M * derjohn well, I'm glad to tell you that there are chatlogs of this conversation ;)
1152715249 M * daniel_hozac for me it just froze my system solid for about a minute.
1152715252 M * FaUl which bug? have i to update all my kernels?
1152715300 M * derjohn daniel_hozac, with which RC was the bug introduced?
1152715312 M * daniel_hozac -rc23 reintroduced it.
1152715323 M * daniel_hozac we fixed it a while back.
1152715331 M * derjohn [PUPPETS]Gonzo, which rc did you use?
1152715343 M * [PUPPETS]Gonzo I used rc17 and rc24
1152715345 M * daniel_hozac -rc25 fixed it again.
1152715363 M * daniel_hozac -rc19 fixed it the first time around.
1152715364 M * FaUl DOH!
1152715368 M * FaUl -rc24
1152715368 M * [PUPPETS]Gonzo yeah
1152715390 M * [PUPPETS]Gonzo derjohn seems to be lucky for picking cool patchlevels *g*
1152715391 M * derjohn http://linux-vserver.org/ChangeLogDevelExperimental
1152715404 M * derjohn vs2.1.1-rc25 fixed it.
1152715434 M * derjohn I always take bleeding stuff, only with real food i am scary (vegetarian) :)
1152715454 M * vrwttnmtu "[PUPPETS]Gonzo warn noone, as people might get to know the bug ;)" : That's not the spirit, is it? :) Security through obscurity?
1152715460 M * [PUPPETS]Gonzo ;)
1152715478 M * [PUPPETS]Gonzo derjohn: stop chatting now, your pc needs all cpu for compiling -rc26 for me ;)
1152715493 M * vrwttnmtu Actually, one of the main problems I have with linux-vserver is that there isn't anywhere that says >= 2.1.1-rcxxxx is secure
1152715509 M * daniel_hozac vrwttnmtu: nothing is secure :)
1152715516 M * vrwttnmtu daniel_hozac, OK, known secure :)
1152715548 M * vrwttnmtu I'd like a page that I can go to that says the latest (known!) secure version of the utils, the patch, and the linux kernel
1152715562 M * daniel_hozac the latest is always the most secure.
1152715568 M * vrwttnmtu :)
1152715577 M * daniel_hozac from a "known vulnerabilities" standpoint.
1152715582 M * vrwttnmtu OK, the earliest version with no known security vulns
1152715582 M * [PUPPETS]Gonzo well, not with -rc17 and -rc23 ;)
1152715590 M * [PUPPETS]Gonzo k
1152715597 M * vrwttnmtu Bugs are one thing
1152715601 M * vrwttnmtu Security bugs are another
1152715624 M * daniel_hozac 2.1 gets new features/rewrites all the time.
1152715625 M * vrwttnmtu So rc22 is vulnerable?
1152715629 M * daniel_hozac no.
1152715666 M * daniel_hozac AFAICT, -rc9 to -rc18 and -rc23 to -rc24 are vulnerable.
1152715669 M * vrwttnmtu daniel_hozac, 2.1 might get new versions all the time, but Bombom's IPv6 patch doesn't
1152715671 M * vrwttnmtu :)
1152715705 M * daniel_hozac vrwttnmtu: i'm using it on 2.6.17.4-vs2.1.1-rc25 right now ;)
1152715730 A * derjohn presses thumbs that we'll get a release soon
1152715761 M * daniel_hozac derjohn: Bertl_oO said "PLM, code review, PLM, then..." ;)
1152715774 M * derjohn PLM?
1152715808 M * daniel_hozac OSDL's kernel compile testing thingie.
1152715819 M * daniel_hozac http://plm.osdl.org/plm-cgi/plm?module=home
1152715825 M * derjohn yeah, I know
1152715831 M * derjohn but didnt know the na,e
1152715874 M * vrwttnmtu http://linux-vserver.org/Security <-- New page
1152715926 M * daniel_hozac vrwttnmtu: i'd say 2.6.17.4 and -rc26.
1152715940 M * daniel_hozac 2.6.16.20 is 3 or 4 releases old by now.
1152715989 M * vrwttnmtu daniel_hozac, But if there aren't any security vulnerabilities in it....
1152715994 M * daniel_hozac and i thought Gentoo was at 0.30.210-r14 :)
1152716003 M * vrwttnmtu daniel_hozac, But if there aren't any security vulnerabilities in it....
1152716004 M * daniel_hozac vrwttnmtu: there are. that's why they released the updates :)
1152716011 M * vrwttnmtu really?
1152716021 M * vrwttnmtu There aren't any GLSA's for the utils
1152716025 M * daniel_hozac for 2.6.16, definitely. read the ChangeLogs.
1152716038 M * daniel_hozac not for the utils. those were just functionality fixes, i think.
1152716038 J * gerrit ~gerrit@67.160.146.170
1152716046 M * vrwttnmtu It really annoys me that they stopped doing kernel GLSAs
1152716060 M * vrwttnmtu It was a good method of tracking bugs
1152716070 M * vrwttnmtu daniel_hozac, Are they local root, DoS, what?
1152716078 Q * newz2000 Quit: Talk to you later
1152716144 M * daniel_hozac i don't remember, i haven't tracked 2.6.16 since .20.
1152716160 M * daniel_hozac seems to be a slice of everything, though.
1152716408 M * daniel_hozac vrwttnmtu: btw, http://daniel.hozac.com/vserver/util-vserver-0.30.210-ipv6.patch in case you weren't aware of that.
1152716472 M * vrwttnmtu What does that do, Mr Danien-san?
1152716487 Q * TermUnitX Read error: Connection reset by peer
1152716501 M * daniel_hozac IPv6 support for the utils.
1152716514 M * vrwttnmtu I guessed that :)
1152716518 M * vrwttnmtu But it all works without it
1152716526 M * daniel_hozac i.e. echo 2001::1/64 > /etc/vservers/<name>/interfaces/0/ip
1152716545 M * vrwttnmtu Aaah
1152716598 M * daniel_hozac it requires that you have chbind6 in your $PATH though.
1152716617 M * vrwttnmtu OK
1152716665 M * vrwttnmtu Can't you get those patches in the main version? Bribe Bertl or something?
1152716666 M * vrwttnmtu :)
1152716679 M * daniel_hozac Bertl_oO isn't the one maintaining the tools.
1152716707 M * vrwttnmtu Well, the main vserver patch, the tools, everything. Who does maintain the tools then?
1152716717 M * daniel_hozac Enrico.
1152716804 M * vrwttnmtu I don't think I've ever seen him speak on this channel. /me waves at Enrico
1152716816 M * daniel_hozac he hasn't been on this channel on a regular basis for months.
1152716832 J * TermUnitX ~PhAnATiC@201.129.207.251
1152716909 N * otaku42 otaku42_away
1152717468 M * daniel_hozac anonc: are you able to test the link breaking on chown/chmod/utimes stuff? (i think i got the dentry ref counting right now)
1152717621 J * Viper0482 ~Viper0482@p54976B25.dip.t-dialin.net
1152717794 Q * vrwttnmtu Quit: Leaving
1152718388 J * stefani ~stefani@tsipoor.banerian.org
1152719523 Q * sladen Quit: bye bye
1152719908 Q * TermUnitX Quit:  (-(PS)-) [v5.0.r02]  http://www.kalendas.net
1152720399 Q * gerrit Ping timeout: 480 seconds
1152720451 Q * ||Cobra|| Remote host closed the connection
1152721041 J * bonbons ~bonbons@83.222.39.166
1152721397 J * FuraX ~FuraX@umb-sls99-003.u-strasbg.fr
1152721852 M * Skram oy! i sitll havent started my presentation VServer!
1152722248 M * Skram Bertl / Anyone.. where is the Linux-VServer logo?
1152722379 M * daniel_hozac http://vserver.13thfloor.at/Stuff/LOGO/Linux-VServer-1.02.eps
1152722383 Q * schimmi Ping timeout: 480 seconds
1152724040 M * AndrewLee waldi: I found CONFIG_VSERVER_HARDCPU is not set in linux-image-2.6.17-1-vserver-686, any reason not set it as default for debian stock kernel?
1152724134 M * AndrewLee waldi: Without it enable, I cannot use sched_prio flag
1152724156 M * AndrewLee waldi: s/sched_prio/sched_hard/
1152724270 J * gerrit ~gerrit@bi01p1.co.us.ibm.com
1152725019 M * waldi hmm
1152725022 M * waldi have to check
1152725049 J * schimmi ~sts@212.202.73.176
1152725239 M * waldi +       depends on EXPERIMENTAL
1152725239 M * waldi +       default n
1152725453 M * AndrewLee waldi: Could you make a exception on it?
1152725504 M * waldi have to think about
1152726146 M * AndrewLee waldi: cause if user choose the -vserver stock kernel on a server, most of them will need sched_hard flag. If the stock -vserver kernel doesn't support, they will still keep rebuild kernel from source.
1152726181 M * AndrewLee waldi: And that make -vserver stock kernel useless for server users.
1152726743 J * s0undt3c1 ~s0undt3ch@bl7-244-155.dsl.telepac.pt
1152727096 J * clessing ~clessing@87.160.230.71
1152727148 Q * s0undt3ch Ping timeout: 480 seconds
1152727154 N * s0undt3c1 s0undt3ch
1152727290 P * clessing 
1152727702 M * h01ger to make a vserver not started by default (anymore) i just remove /etc/vservers/$name/apps/init ?
1152727729 M * daniel_hozac +/style, yes.
1152727730 J * matti_ ~matti@linux.gentoo.pl
1152727739 M * doener s/style/mark/
1152727757 M * daniel_hozac uh, right.
1152727768 A * daniel_hozac should've learned not to watch TV and type at the same time.
1152727794 M * h01ger not s/mark/h01ger/ ? :)
1152727828 M * doener I don't think that util-vserver cares about h01ger :)
1152727903 Q * matti Ping timeout: 480 seconds
1152727903 N * matti_ matti
1152727916 M * h01ger oh well... 
1152727937 A * h01ger should have learned not to do so many things at the same time.
1152728608 J * vrwttnmtu ~eryktyktu@82-69-161-137.dsl.in-addr.zen.co.uk
1152728615 M * vrwttnmtu Bonjour
1152729465 J * shedi ~siggi@130.208.221.254
1152730836 M * Skram where is the Linux-VServer LOGO
1152730855 M * daniel_hozac Skram: i told you the last time you asked.
1152730869 M * Skram daniel_hozac: I didnt see the link
1152730870 M * Skram sorry
1152730939 M * mnemoc [18:38:36] <Skram> Bertl / Anyone.. where is the Linux-VServer logo?
1152730940 M * mnemoc [18:40:47] <daniel_hozac> http://vserver.13thfloor.at/Stuff/LOGO/Linux-VServer-1.02.eps
1152730945 M * Skram thank you
1152731932 P * meandtheshell 
1152732398 Q * shedi Quit: Leaving
1152733270 J * SiD4WiNDR luser@195.160.166.163
1152733327 Q * sid3windr Quit: In Soviet Russia, the quit messages you!
1152733337 N * SiD4WiNDR sid3windr
1152733342 J * Smutje_ ~Smutje@xdsl-87-78-2-222.netcologne.de
1152733454 Q * Smutje Ping timeout: 480 seconds
1152733454 N * Smutje_ Smutje
1152733601 P * stefani I'm Parting (the water)
1152733823 M * vrwttnmtu Quit: In Soviet Russia, the quit messages you! <-- Heh
1152733828 Q * nox Ping timeout: 480 seconds
1152733844 M * vrwttnmtu Or the messages quit you?
1152733909 Q * lehkor_ Quit: leaving
1152733975 J * matti_ ~matti@212.244.232.46
1152734011 M * vrwttnmtu daniel_hozac, If you're around, I've re-emerged gcc, binutils, and glibc, and then recompiled chbind6, and it still doesn't work :(
1152734041 M * daniel_hozac you'll have to recompile util-vserver.
1152734045 M * vrwttnmtu Aaah, OK
1152734048 M * vrwttnmtu 2 secs
1152734109 M * vrwttnmtu Love the addition of the VDIRBASE env variable.
1152734118 Q * mire Quit: Leaving
1152734129 Q * matti Ping timeout: 480 seconds
1152734129 N * matti_ matti
1152734137 M * vrwttnmtu (Maybe it was there before, and I didn't notice it, but either way)
1152734150 N * matti Guest483
1152734155 M * vrwttnmtu So the util-vserver includes a library that chbind6 uses?
1152734161 M * daniel_hozac yes.
1152734171 M * vrwttnmtu Will I need to recompile chbind6 after this?
1152734179 M * vrwttnmtu Or will it just use the new lib?
1152734180 M * daniel_hozac no, it links it dynamically.
1152734186 M * vrwttnmtu Hokay
1152734274 M * vrwttnmtu :( It still doesn't work :(
1152734337 M * vrwttnmtu I tried copying another chbind from another host, and it didn't work
1152734350 M * daniel_hozac try copying the libvserver.so.0.0.0 instead.
1152734396 M * vrwttnmtu The most likely thing I can think of is that this kernel is compiled with less stuff in it than the others
1152734405 M * vrwttnmtu For eg: <M> PF_KEY sockets
1152734413 M * vrwttnmtu Not compiled in statically
1152734433 M * vrwttnmtu <*>   The IPv6 protocol
1152734461 M * daniel_hozac that really shouldn't make the syscall fail sporadically.
1152734516 M * vrwttnmtu In my vserver options I have: Legacy Kernel API, dynamic context IDs, COW, Proc Security, Hard CPU, and nothing else
1152734582 A * vrwttnmtu doesn't get this.
1152734751 M * vrwttnmtu The util-vserver on all 3 boxes is the same
1152734755 M * vrwttnmtu version I mean
1152734768 M * daniel_hozac vrwttnmtu: get http://vserver.13thfloor.at/Experimental/SYSCALL/syscall_shiny10.h and then http://daniel.hozac.com/vserver/chbind6-shiny.c
1152734787 M * vrwttnmtu OK
1152734847 M * vrwttnmtu I assume I compile, and use that header
1152734856 M * daniel_hozac the header is used automatically.
1152734869 M * vrwttnmtu OK, so use the same compile command?
1152734874 M * daniel_hozac yeah.
1152734891 M * vrwttnmtu chbind6-shiny.c: In function `vserver':
1152734891 M * vrwttnmtu chbind6-shiny.c:30: error: `__NR_vserver' undeclared (first use in this function)
1152734909 M * daniel_hozac whoops, add a -D__NR_vserver=273 (if you're on x86).
1152734966 M * vrwttnmtu It worked!
1152734970 M * vrwttnmtu I love you!
1152734974 M * vrwttnmtu What was it?
1152734988 M * daniel_hozac that is _extremely_ odd.
1152735001 M * vrwttnmtu !!**EXREMELY**!! ? :)
1152735020 M * daniel_hozac pretty much.
1152735056 J * nox ~nox@noxlux.de
1152735066 M * daniel_hozac that should be the exact same code that is executed by the original one.
1152735089 M * vrwttnmtu # diff chbind6.c chbind6-shiny.c
1152735089 M * vrwttnmtu 15a16
1152735089 M * vrwttnmtu > #include "syscall_shiny10.h"
1152735089 M * vrwttnmtu 28a30,31
1152735089 M * vrwttnmtu > _syscall3(int, vserver, uint32_t, cmd, uint32_t, id, void *, data);
1152735090 M * vrwttnmtu >
1152735092 M * vrwttnmtu 106c109,111
1152735094 M * vrwttnmtu <               vc_syscall(VCMD_net_add, nid, &addrs[addr-1]);
1152735096 M * vrwttnmtu ---
1152735098 M * vrwttnmtu >               if (vserver(VCMD_net_add, nid, &addrs[addr-1]) == -1) {
1152735100 M * vrwttnmtu >                       fprintf(stderr, "chbind6: vc_net_add(%d): %s\n", addr-1, strerror(errno));
1152735102 M * vrwttnmtu >               }
1152735119 M * daniel_hozac do you use any odd compiler flags on that host?
1152735128 M * daniel_hozac (for emerge, i mean)
1152735132 M * vrwttnmtu CFLAGS="-O2 -march=pentium4 -fomit-frame-pointer"
1152735177 M * daniel_hozac what happens if you compile the new chbind6 with those flags?
1152735217 M * vrwttnmtu How can I remove the v6 addr?
1152735245 M * vrwttnmtu Is there a -d for chbind6 ?
1152735260 M * vrwttnmtu Or do I just remove it from the host i/f and re-add it?
1152735273 M * mnemoc vrwttnmtu: diff -u 
1152735284 Q * coocoon Ping timeout: 480 seconds
1152735288 M * vrwttnmtu Diffing a binary?
1152735356 M * daniel_hozac chbind6 doesn't have a -d yet.
1152735374 M * mnemoc vrwttnmtu: no, in general... when diffing please use -u
1152735385 M * vrwttnmtu Oh :)
1152735388 M * vrwttnmtu Sorry.
1152735658 M * vrwttnmtu daniel_hozac, the shiny one works when compiled with those CFLAGS
1152735675 M * vrwttnmtu # ping6 www.kame.net
1152735675 M * vrwttnmtu PING www.kame.net(orange.kame.net) 56 data bytes
1152735675 M * vrwttnmtu 64 bytes from orange.kame.net: icmp_seq=1 ttl=49 time=276 ms
1152735675 M * vrwttnmtu --- www.kame.net ping statistics ---
1152735675 M * vrwttnmtu 1 packets transmitted, 1 received, 0% packet loss, time 0ms
1152735676 M * vrwttnmtu rtt min/avg/max/mdev = 276.116/276.116/276.116/0.000 ms
1152735682 M * vrwttnmtu Wahoo!
1152735684 M * vrwttnmtu :
1152735739 M * vrwttnmtu cat /dev/random > chbind6.c && gcc chbind6.c && ./a.out -a 2001:1/64
1152735959 M * daniel_hozac vrwttnmtu: http://daniel.hozac.com/vserver/chbind6-shiny.c chbind6 -n ... -d ... should delete addresses.
1152735961 J * coocoon ~coocoon@84.160.95.55
1152735980 M * vrwttnmtu daniel_hozac, It's OK, I added another one
1152736125 M * vrwttnmtu Well, I wonder what the difference is.
1152736471 Q * Viper0482 Quit: one day, i'll find this peer guy and then i'll reset his connection!!
1152737593 N * Ben_ Ben_zZz
1152738111 M * vrwttnmtu ./chbind6-shiny: invalid option -- d
1152738112 M * vrwttnmtu Hmm.
1152738114 M * daniel_hozac vrwttnmtu: i updated your instructions to match the new chbind6.
1152738124 M * vrwttnmtu On the IPv6 page?
1152738127 M * vrwttnmtu Wicked
1152738159 M * vrwttnmtu Apache is bound to my vserver ipv6 address, but postfix doesn't see it.
1152738163 M * vrwttnmtu I've no idea why
1152738182 M * daniel_hozac postfix has options regarding which protocols to bind.
1152738192 M * vrwttnmtu Really?
1152738197 M * vrwttnmtu Must be a new change?
1152738228 M * vrwttnmtu Master .cf?
1152738235 M * daniel_hozac main.cf IIRC.
1152738257 M * daniel_hozac inet_protocols.
1152738280 M * vrwttnmtu inet_protocols = ipv4, ipv6 (enable both IPv4 and IPv6)
1152738281 M * vrwttnmtu Yep
1152738283 M * vrwttnmtu Just found it
1152738330 M * vrwttnmtu tcp        0      0 :::25                   :::*                    LISTEN      0          1222424    22290/master
1152738331 M * vrwttnmtu Lovely
1152738674 J * sladen paul@starsky.19inch.net
1152739328 Q * bonbons Quit: Leaving
1152739788 Q * yarihm Quit: Leaving
1152739801 J * icarus hiddenserv@tor.noreply.org
1152739898 Q * sladen Ping timeout: 480 seconds
1152739913 Q * coocoon Quit: KVIrc 3.2.0 'Realia'
1152740054 J * shedi ~siggi@130.208.221.254
1152740152 J * sladen paul@starsky.19inch.net
1152740317 J * Blissex ~Blissex@82-69-39-138.dsl.in-addr.zen.co.uk
1152740427 Q * icarus Quit: Leaving
1152740542 J * icarus hiddenserv@tor.noreply.org
1152741090 J * Aiken ~james@tooax6-122.dialup.optusnet.com.au
1152741260 Q * Guest483 Read error: Connection reset by peer
1152741312 J * matti ~matti@212.244.232.46
1152741872 Q * Blissex Remote host closed the connection
1152742005 Q * icarus Quit: Leaving
1152743889 Q * romke Remote host closed the connection
1152743922 J * matti_ ~matti@linux.gentoo.pl
1152743963 J * romke ~romke@83.16.133.162
1152743979 Q * matti Read error: Connection reset by peer
1152744799 Q * shedi Read error: Connection reset by peer
1152745032 Q * Aiken Quit: Leaving
1152745188 J * Aiken ~james@tooax6-122.dialup.optusnet.com.au
1152745677 J * shedi ~siggi@130.208.221.254
1152745740 J * mire_ ~mire@156-166-222-85.COOL.ADSL.VLine.verat.net
1152746463 Q * dna Quit: Verlassend
1152746825 Q * mire_ Quit: Leaving
1152747228 Q * vrwttnmtu Quit: Leaving