1274403106 Q * petzsch Quit: Leaving.
1274404156 Q * infowolfe Quit: Leaving
1274404313 J * derjohn_foo ~aj@f048106250.adsl.alicedsl.de
1274404753 Q * derjohn_mob Ping timeout: 480 seconds
1274405214 J * infowolfe ~infowolfe@c-71-236-152-35.hsd1.or.comcast.net
1274410864 J * SauLus_ ~SauLus@80.171.26.95
1274411270 Q * SauLus Ping timeout: 480 seconds
1274411270 N * SauLus_ SauLus
1274416963 J * sharkjaw ~gab@90.149.128.29
1274419719 Q * FireEgl Remote host closed the connection
1274420531 J * FireEgl Proteus@2001:470:e056:1:223:54ff:fe89:b207
1274420694 J * mtg ~mtg@port-87-193-189-26.static.qsc.de
1274420862 J * ghislain ~AQUEOS@adsl2.aqueos.com
1274421100 Q * FireEgl Read error: Connection reset by peer
1274421264 J * ncopa ~ncopa@180.40.189.109.customer.cdi.no
1274421436 Q * derjohn_foo Ping timeout: 480 seconds
1274421913 J * FireEgl FireEgl@173-16-9-10.client.mchsi.com
1274422147 J * ghislain1 ~AQUEOS@adsl2.aqueos.com
1274422364 Q * ghislain Ping timeout: 480 seconds
1274422430 J * awk ~awk@gw1.security.web.za
1274422435 M * awk helloooo :D
1274422462 M * awk Wondering if anyone has a quick fix.. 
1274422463 M * awk May 20 21:09:01 pbx crond[7962]: CRON (scopserv) ERROR: failed to open PAM security session: Success .. May 20 21:09:01 pbx crond[7962]: CRON (scopserv) ERROR: cannot set security context ...
1274422481 M * awk According to this error, it looks like its related to /etc/shadow .. and this is missing on my vserver
1274422490 J * imcsk8 ~ichavero@189.155.226.165
1274422508 M * awk I installed auethconfig and ran autoconfig --enableshadow which doesn't seem to do work
1274422515 M * awk Not sure what is in the vserver is messing with cron
1274423532 J * derjohn_foo ~aj@139.12.1.252
1274423696 J * awk_ ~awk@gw1.security.web.za
1274423724 M * awk_ erp, disconnected, anyoe manage to answer my question?
1274423733 Q * awk Ping timeout: 480 seconds
1274423740 N * awk_ awk
1274423823 Q * imcsk8 Quit: This computer has gone to sleep
1274424392 M * Swen awk: i have a sloution
1274424407 M * Swen is this a CentOS guest?
1274424544 M * awk access.conf and added + : ALL : cron crond also I checked pam.d crond config and looks correct
1274424548 M * awk Swen yup
1274424552 M * Swen look at:  http://linux-vserver.org/Frequently_Asked_Questions#Why_do_neither_sshd_nor_crond_.28vixie-cron.29_work_correctly_in_my_CentOS_.2F_Fedora_guest.3F_I_get_.27pam_loginuid.28crond:session.29:_set_loginuid_failed_opening_loginuid.27_and_similar_lines_in_my_logs.
1274424719 M * awk woop woop :) thanks! let me try that.
1274424749 M * Swen np
1274425213 J * ntrs ~ntrs@77.29.115.142
1274426337 J * BenG ~bengreen@cpc2-aztw22-2-0-cust521.aztw.cable.virginmedia.com
1274426364 Q * BenG 
1274427002 M * awk yay, that works nicely! :D
1274427896 Q * saros Quit: leaving
1274427927 Q * Piet Remote host closed the connection
1274427984 J * Piet ~Piet__@04ZAACJGP.tor-irc.dnsbl.oftc.net
1274428315 J * barismetin ~barismeti@zanzibar.inria.fr
1274428881 J * ntrs_ ~ntrs@77.28.28.133
1274428882 Q * ntrs Read error: Connection reset by peer
1274429288 J * thierryp ~thierry@zankai.inria.fr
1274429463 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d
1274429790 J * gnuk ~F404ror@pla93-3-82-240-11-251.fbx.proxad.net
1274431844 J * BenG ~bengreen@cpc2-aztw22-2-0-cust521.aztw.cable.virginmedia.com
1274432088 N * Bertl_zZ Bertl
1274432099 M * Bertl morning folks!
1274432548 M * hijacker morning
1274432698 M * awk Morning Bertl: please note, I never required your help today, Swen pulled one out of the hat!
1274432699 M * awk :D
1274432707 M * awk sneeky little crond
1274433521 J * selim ~chatzilla@e181067004.adsl.alicedsl.de
1274433526 P * selim 
1274433714 M * Bertl awk: fascinating :)
1274434781 Q * waldgeist Ping timeout: 480 seconds
1274434827 J * petzsch ~markus@dfn371.rz.tu-ilmenau.de
1274434982 J * richard_ ~richard@p50902DF1.dip.t-dialin.net
1274434984 M * Swen awk: it's just because i've read all documentation on Vserver several times :D
1274434994 J * waldgeist ~ignaz@85-126-150-194.work.xdsl-line.inode.at
1274435000 M * richard_ hallo
1274435100 M * petzsch richard_: hi (english please)
1274435145 M * Bertl richard_: that means: hello :)
1274435208 M * richard_ ah sorry
1274435225 M * richard_ i need some help
1274435241 M * richard_ vserver on debian is not a problem, but on opensuse.
1274435284 M * richard_ i will post the output on a pastebin site - need some time - damn touchpad...
1274435333 M * petzsch one more distribution for Berl's bad with vserver list *g*
1274435416 M * Bertl either nobody uses opensuse or it wasn't problematic till now .. at least I haven't got any bad feedback yet
1274435521 M * richard_ the most time i use debian but on my laptop i use opensuse and i need a debian guest
1274435531 M * richard_ http://pastebin.com/tsnYky4Q
1274435557 M * richard_ i have patched the kernel and compiled the util-vserver and installed debootstrap 0.3.x
1274435570 M * Bertl please upload the output from:
1274435575 M * richard_ hmm... i need a mouse...
1274435578 M * Bertl vserver-info - SYSINFO
1274435583 M * richard_ okay
1274435736 Q * BenG Quit: I Leave
1274435846 M * richard_ http://pastebin.com/HaXhSK5r
1274435899 M * Bertl Use dietlibc: no (you have been warned)
1274435901 M * daniel_hozac build util-vserver with dietlibc, or install a recent version of it.
1274435928 M * richard_ okay
1274435929 M * Bertl richard_: you want a newer version anyway for that kernel
1274435942 M * Bertl i.e. the 0.30.215 will not provide proper isolation for it
1274435957 M * richard_ is 0.30.216 out?
1274436243 M * Bertl not that I'd know ... but there are plenty of pre releases, while 0.30.215 is more than two years old
1274436244 M * richard_ btw the patch for 2.6.34 has some problems with fs/namei.o
1274436259 M * Bertl you need to disable CoW link breaking for now
1274436290 M * Bertl btw, daniel_hozac, did you find the time to take a look at it?
1274436375 M * richard_ good to know, but 2.6.33.3 is also okay :)
1274436388 M * Bertl I hope so :)
1274436541 M * richard_ never try to compile a kernel on a laptop, it needs damn much time...
1274436686 M * richard_ hm, now i got a error at compiling with dietlibc
1274436708 M * Bertl depends on the kernel config, and what dietlibc error?
1274436828 M * richard_ http://pastebin.com/xWTAK8jn
1274436941 M * Bertl what dietlibc version is that?
1274437045 M * richard_ 0.31
1274437085 J * BenG ~bengreen@cpc2-aztw22-2-0-cust521.aztw.cable.virginmedia.com
1274437137 M * Bertl that should be fine, maybe a gcc problem?
1274437152 Q * BenG 
1274437198 M * Bertl gcc 4.4.0?
1274437199 M * waldgeist maybe? i will play arround with the versions...
1274437215 M * waldgeist yep
1274437219 M * waldgeist 4.4.0
1274437221 M * Bertl at least google lists that as known regression
1274437227 Q * Mr_Smoke Read error: Connection reset by peer
1274437238 M * Bertl i.e. get a working gcc and you should be fine :)
1274437259 J * Mr_Smoke ~smokey@layla.lecoyote.org
1274437346 Q * petzsch Quit: Leaving.
1274437378 M * Bertl alternatively you can edit the stdarg.h
1274437388 M * Bertl and change #define va_start(v, l)         __bultin_stdarg_start((v), l)
1274437396 M * Bertl to
1274437407 M * Bertl #define va_start(v, l)         __bultin_va_start((v), l)
1274437433 M * waldgeist okay
1274438039 M * richard_ so thax for the help, i will try it a little bit later :)
1274438164 M * Bertl np, you're welcome!
1274439986 J * JonB ~NoSuchUse@130.227.63.19
1274440040 M * JonB hey Bertl 
1274440132 M * Bertl hey JonB!
1274440174 M * JonB Bertl: i'm gonna give a little talk about linux vserver tuesday. Any new recent development in the time i have not been following so closely?
1274440186 M * JonB probably 1-2 years
1274440192 M * Bertl lol
1274440207 M * Bertl well, we are using a lot of mainline stuff by now
1274440216 M * JonB Bertl: well, i have been using it daily, but no new kernels
1274440222 M * Bertl (at least the parts which are useable)
1274440245 M * Bertl a bunch of new ccaps and cflags have been added
1274440254 M * JonB sounds good
1274440272 M * JonB any security monitoring for changes to files?
1274440278 M * JonB changes inside the guest
1274440300 M * Bertl nothing Linux-VServer specific
1274440305 M * JonB ok
1274440322 M * daniel_hozac you already have inotify...
1274440330 M * JonB any other controlling than the console based?
1274440378 M * daniel_hozac Bertl: http://people.linux-vserver.org/~dhozac/p/k/delta-cow-fix22.diff
1274440384 M * Bertl JonB: hmm?
1274440387 M * daniel_hozac not tested yet, but it compiles.
1274440393 M * Bertl daniel_hozac: ah, great!
1274440398 M * JonB Bertl: some webthing perhaps?
1274440416 M * Bertl tux is the only web server inside the kernel for now :)
1274440417 M * daniel_hozac i will go over it more closely later.
1274440441 M * JonB Bertl: i ment like the util-vserver packages
1274440539 M * Bertl well, there are a bunch of control panels around Linux-VServer 
1274440553 M * Bertl but I think the is outside the scope of Linux-VServer itself
1274440573 M * JonB i agree, but i do want to mention that they exist
1274441308 Q * ntrs_ Ping timeout: 480 seconds
1274443211 J * ntrs ~ntrs@77.28.173.3
1274446087 J * petzsch ~markus@dslb-094-222-102-026.pools.arcor-ip.net
1274447423 Q * sharkjaw Remote host closed the connection
1274447527 J * aimbot moon@40.net219096018.t-com.ne.jp
1274447530 M * aimbot Hello
1274447619 M * aimbot i am trying to follow http://linux-vserver.org/Installation_on_CentOS guid to install vserver on my CentOS 4.8, in the step where i have to add dhozac-vserver.repo lines, i add it, then i type 'yum update yum' and... nothing happens, also when i type 'yum install kernel' - nothing happens ;(
1274447772 M * Bertl maybe daniel_hozac knows .. but you can always install from sources
1274447830 M * aimbot i preffer not to
1274447837 M * aimbot im doing it on remote system :p
1274447871 M * Bertl I usually build my kernels (and tools) on remote systems, so?
1274447899 M * aimbot i dont have config needed
1274447910 M * aimbot so i would have to select everything in the kernel
1274447911 M * aimbot :D
1274447946 M * Bertl so you better off with a precompiled version which selected everything in the kernel?
1274447977 M * aimbot hmpf:p
1274447987 M * aimbot well, if nothing else works i will do from sources
1274447988 M * Bertl anyway, it's your choice ... I'm just wondering about the argument
1274448009 M * aimbot but for now i want to do it via yum: 1. its faster, 2. it's easier ;)
1274448146 M * aimbot is there other centos repo from where i can get the vserver kernel
1274448150 M * aimbot darn
1274448179 M * Bertl I doubt so, but there might be a mirror
1274448189 M * Bertl (not sure how that will help though)
1274448486 M * aimbot Note that since late December, the dhozac repo has frequently been unavailable. If this is the case then the 
1274448489 M * aimbot ha !
1274448511 M * aimbot found it on http://wiki.centos.org/HowTos/Virtualization/Vserver (in case someone needs it), seems it solved my repo problem :)
1274448846 M * Bertl excellent!
1274448873 M * Bertl maybe link it from the Linux-VServer wiki (centos page)
1274449335 M * daniel_hozac do you have yum priorities installed?
1274449341 M * daniel_hozac you'll have to disable that.
1274449547 M * aimbot daniel_hozac: 
1274449572 M * aimbot when im using your repo on my centos 4.8 it fail to install, when i manually set i have centos 5 then it seems to work (the kernel installing)
1274449596 M * aimbot look
1274449607 M * aimbot http://rpm.hozac.com/dhozac/centos/4/vserver/i386/ <- here is no vserver kernel
1274449616 M * aimbot http://rpm.hozac.com/dhozac/centos/5/vserver/i386/ <- and here it is
1274449617 M * aimbot :/
1274449693 M * aimbot it looks like for 4.x branch some rpm files are missing...
1274449716 M * thierryp Bertl: hi
1274449738 M * thierryp just asking, are you guys providinga git feed ?
1274449872 Q * petzsch Quit: Leaving.
1274449873 M * aimbot i'm unable to finish installing vserver with these rpm files missing, unless i go for manual compilation which is very slow..
1274450108 M * Bertl thierryp: no
1274450378 M * daniel_hozac aimbot: yeah, CentOS 4 is not supported.
1274450390 M * aimbot oh damn
1274450408 M * aimbot so what i cna do? only manual kernel compilation will make me vserver working on c4 ?
1274450420 M * aimbot or i have to switch to centos 5 ?
1274450508 M * daniel_hozac manual would probably work...
1274451730 N * Bertl Bertl_zZ
1274452038 Q * JonB Read error: Connection reset by peer
1274452408 Q * FireEgl Read error: Operation timed out
1274453105 J * Walex ~Walex@82-69-39-138.dsl.in-addr.zen.co.uk
1274453259 J * dowdle ~dowdle@scott.coe.montana.edu
1274453420 J * imcsk8 ~ichavero@evdomip-48-72.iusacell.net
1274453578 Q * thierryp Quit: ciao folks
1274453831 Q * imcsk8 Quit: This computer has gone to sleep
1274454231 Q * derjohn_foo Ping timeout: 480 seconds
1274455171 J * FireEgl Proteus@2001:470:e056:1:223:54ff:fe89:b207
1274456242 N * Bertl_zZ Bertl
1274456448 Q * ncopa Quit: Ex-Chat
1274457743 Q * mtg Quit: Verlassend
1274458750 J * thierryp ~thierry@home.parmentelat.net
1274458949 J * ntrs_ ~ntrs@77.29.114.233
1274459191 Q * richard_ Remote host closed the connection
1274459355 Q * Walex Ping timeout: 480 seconds
1274459385 Q * ntrs Ping timeout: 480 seconds
1274459512 Q * barismetin Remote host closed the connection
1274460711 Q * FireEgl Remote host closed the connection
1274461708 J * FireEgl Proteus@2001:470:e056:1:223:54ff:fe89:b207
1274462320 Q * FireEgl Ping timeout: 480 seconds
1274462894 J * FireEgl FireEgl@173-16-9-10.client.mchsi.com
1274463871 J * hijacker_ ~hijacker@87-126-142-51.btc-net.bg
1274464241 Q * gnuk Quit: NoFeature
1274466460 Q * ghislain1 Quit: Leaving.
1274466845 M * fzylogic any ETA on a 2.6.34 patch?
1274466884 M * daniel_hozac it's on http://vserver.13thfloor.at/Experimental/
1274466901 M * fzylogic the pre2 patch didn't compile for me
1274466911 M * daniel_hozac disable COW.
1274466920 M * daniel_hozac or, try it with http://people.linux-vserver.org/~dhozac/p/k/delta-cow-fix22.diff as well.
1274466926 M * fzylogic ok
1274467042 M * fzylogic turns out my bug with random processes being set to OOM_DISABLE exists in mainline as well, but it's far more prevalent on vservers
1274467093 Q * thierryp Ping timeout: 480 seconds
1274467094 M * fzylogic want to try 2.6.34 before reporting it to lkml
1274467196 M * fzylogic and I still haven't found any bugs in ipv6 so good job with that  :)
1274467725 M * fzylogic kernel BUG at kernel/vserver/network.c:182!
1274467752 M * fzylogic about 2 seconds after booting completes
1274467819 M * fzylogic full trace:
1274467819 M * fzylogic http://paste.linux-vserver.org/15334
1274467866 J * petzsch ~markus@p4FF468A2.dip.t-dialin.net
1274468335 M * daniel_hozac without doing anything vserver related?
1274468401 M * fzylogic since util-vserver backgrounds the guest startup, I can't easily guarantee that
1274468406 M * fzylogic the machine does have a single guest on it
1274468414 M * fzylogic should I try rebooting with that guest disabled?
1274468429 M * daniel_hozac yes.
1274468431 M * daniel_hozac please.
1274468657 M * fzylogic yeah, it's fine with that guest disabled
1274468682 M * fzylogic crashes the moment the guest is started
1274468701 M * Bertl kernel dump?
1274468752 M * fzylogic http://paste.linux-vserver.org/15335
1274468757 M * fzylogic same as the first
1274468902 M * Bertl you are not using network namespaces, right?
1274468966 M * fzylogic no
1274468982 M * fzylogic # CONFIG_NET_NS is not set
1274469143 M * Bertl and it happens when the guest starts?
1274469311 M * fzylogic yeah
1274469320 M * fzylogic before any output is displayed from the vserver command
1274469406 M * Bertl hmm, could it be that the guest doesn't keep running?
1274469420 M * fzylogic the guest starts fine with 2.6.33
1274469423 M * Bertl i.e. that it has no persistant service enabled?
1274469443 M * fzylogic no, it has quite a few services starting by default
1274469449 M * fzylogic ssh, ftp, apache...
1274469484 M * Bertl hmm, well, the function you are hitting indictes that a task exits and the network context is destroyed
1274469489 J * Piet_ ~Piet__@04ZAACJT6.tor-irc.dnsbl.oftc.net
1274469515 M * Bertl could be that the clone part failed to take a reference on it (will check that shortly)
1274469530 M * fzylogic ok
1274469555 M * fzylogic just let me know if I can test anything else.  I have a rare opportunity to test with an otherwise unused machine  :)
1274469615 J * derjohn_foo ~aj@g228050188.adsl.alicedsl.de
1274469668 M * Bertl any limits or oddities for that guest?
1274469697 M * fzylogic memory limit is set to 2300M via cgroups
1274469717 M * fzylogic it's the same default setup we're using on thousands of other instances
1274469798 M * Bertl could you try to narrow the services started on the guest down to the one causing this issue?
1274469824 M * fzylogic if it were a service, wouldn't I see some output from the vserver command?
1274469842 M * Bertl not necessarily
1274469851 Q * Piet Ping timeout: 480 seconds
1274469858 M * Bertl could be the first service started or some preparation
1274469869 M * Bertl I'd suggest to try the following:
1274469884 M * Bertl - clone the guest, see if that causes the same
1274469900 M * Bertl - if yes, try to reduce the active services with chroot
1274469919 M * Bertl - if no, try to figure out the difference in the config
1274469926 M * fzylogic I'll just start removing startup services from /etc/rc2.d
1274469939 M * Bertl make sure to work on a copy
1274469948 M * fzylogic yeah, I'm building a second guest now
1274470026 M * fzylogic well, it crashed while starting that new guest
1274470030 M * fzylogic I guess that's a good sign  :)
1274470035 M * Bertl yep
1274470050 M * fzylogic same place, free_nx_info
1274470100 M * Bertl yes, the nx context is disposed early, before it is properly shut down
1274470145 M * fzylogic I think I'll start by changing the default runlevel  to 1
1274470845 M * fzylogic bit different with default runlevel changed to 1:
1274470846 M * fzylogic http://paste.linux-vserver.org/15336
1274470896 M * Bertl do you have the initial trace?
1274470916 M * Bertl i.e. the one before the free_nx_info
1274470942 M * fzylogic that's the entirety of what was on my serial console
1274470972 M * fzylogic I did end up getting a bit of info in my shell this time, though
1274470986 M * Bertl I presume something was cut off at the top
1274470991 M * fzylogic http://paste.linux-vserver.org/15337
1274471020 M * fzylogic just one line:
1274471021 M * fzylogic [   71.232925] BUG: unable to handle kernel paging request at ffff88ff01c00000
1274471047 M * Bertl but yes, that looks like the cause
1274471082 M * Bertl i.e. something kills chbind (probably indirectly) which causes the issue
1274471319 Q * petzsch Quit: Leaving.
1274471386 J * petzsch ~markus@p4FF468A2.dip.t-dialin.net
1274472065 J * mib_zvnttx 40e71a81@ircip2.mibbit.com
1274472107 M * mib_zvnttx Hey room, whats the best linuc distro to put on a 400MHZ and 320MB RAM
1274472122 M * mib_zvnttx To browse the internet that is
1274472130 M * Bertl a working one which isn't too large :)
1274472152 M * mib_zvnttx Uh huh
1274472156 M * mib_zvnttx Like?
1274472188 M * Bertl like a self compiled one
1274472193 M * fzylogic Bertl: http://paste.linux-vserver.org/15338
1274472204 M * fzylogic triggered via:
1274472205 M * fzylogic vcontext --chroot --xid 25906 --create /bin/bash
1274472221 M * fzylogic then I just ran uptime, ps, ls
1274472227 M * fzylogic and it went boom
1274472233 M * fzylogic no services started at all
1274472245 M * Bertl okay
1274472257 M * fzylogic it also didn't panic the entire system  :)
1274472284 M * fzylogic but perhaps that's what is triggering the nx teardown
1274472289 M * mib_zvnttx Isn' there any ready made distro's?
1274472293 M * mib_zvnttx lol
1274472355 M * Bertl fzylogic: okay, I can reproduce that one!
1274472417 M * PowerKe mib_zvnttx: I've heard some good things about 'Damn Small Linux' or maybe 'Puppy Linux' (didn't try myself though)
1274472434 M * mib_zvnttx Can it browse the net?
1274472450 M * mib_zvnttx Thats all Im looking to do
1274472459 M * mib_zvnttx Just online banking
1274472517 M * PowerKe Check their homepages, damn small linux lists a few browsers
1274472612 M * mib_zvnttx I remember using the LIVECD for Damn Small Linux
1274472622 M * mib_zvnttx But I couldn't find any browsers
1274473078 M * Bertl fzylogic: hmm, yeah, that explains a lot, guess I should have a fix in a few minutes :)
1274473090 M * fzylogic awesome
1274473114 M * Bertl seems I 'lost' the process initialization .. :)
1274473148 M * Bertl no idea how that passed testme.sh, but obviously it did :)
1274473244 M * Bertl http://vserver.13thfloor.at/ExperimentalT/delta-copy-fix01.diff
1274473274 M * Bertl you might also want to check out the CoW fix daniel_hozac uploaded earlier (if you are in the testing mood :)
1274473556 M * fzylogic I'm already running it  :)
1274473628 M * fzylogic oh, in case you missed it, I verified that the random processes getting set OOM_DISABLE is something in mainline that's just more easily triggered under vserver
1274473784 M * Bertl interesting, is a fix available/in work?
1274473804 M * fzylogic that's why I'm testing 2.6.34 now
1274473812 M * fzylogic going to try to reproduce and go to lkml with it if I can
1274473834 M * Bertl i.c. well, good that somebody is testing ... and reporting back!
1274473838 M * Bertl very appreciated!
1274473860 M * fzylogic out of 477 regular servers, only 1 has exhibited the issue, whereas with 169 vserver hosts, a few dozen have shown it
1274473867 M * fzylogic really weird  :)
1274473890 M * Bertl we do a lot more cloning
1274473911 M * Bertl i.e. where normal service use fork() we use clone/unshare
1274473928 M * fzylogic ahhh, that makes a bit more sense
1274473959 M * fzylogic guest started up, but then this:
1274473975 M * fzylogic http://paste.linux-vserver.org/15339
1274474003 M * fzylogic http://paste.linux-vserver.org/15340
1274474018 M * fzylogic I'm heading to lunch.  be back shortly.
1274474025 N * fzylogic fzylogic-lunch
1274474071 M * Bertl that looks like the issue which killed the chbind before :)
1274474090 M * Bertl but it also looks mainline, i.e. not Linux-VServer related at first glance
1274474138 M * Bertl question is, are memory cgroups enabled?
1274474713 Q * hijacker_ Quit: Leaving
1274474848 Q * derjohn_foo Ping timeout: 480 seconds
1274475600 J * richard_ ~richard@p50906406.dip.t-dialin.net
1274475617 M * richard_ hello
1274475627 M * Bertl wb
1274476042 M * richard_ i have tried gcc 4.1 and 3.3 always the same error
1274476068 M * Bertl I doubt that, if you are referring to the vararg builtin
1274476079 M * richard_ yes
1274476088 M * Bertl because it was changed in gcc 4.4
1274476137 M * Bertl so, you might have _installed gcc 4.1 or 3.3, but you definitely used the new toolchain
1274476163 M * Bertl but as I said earlier, you can modify the headers to match the change in gcc
1274476184 M * Bertl (which is, what your distribution should have done :)
1274476303 Q * ntrs_ Ping timeout: 480 seconds
1274476706 N * fzylogic-lunch fzylogic
1274476719 M * fzylogic Bertl: yes, memory cgroups are enabled
1274476759 M * Bertl from the trace, it looks like something inside the guest is trying to access cgroup info?
1274476790 M * Bertl or at least /proc/meminfo, I guess
1274476989 M * fzylogic yeah, we have a guest process that'll poke at /proc/meminfo briefly
1274476997 M * fzylogic let me disable it and see it fixes things
1274477043 M * Bertl trying to recreate it here
1274477103 Q * FireEgl Remote host closed the connection
1274477112 M * fzylogic got no crash with it disabled, but vserver-stat returns the following and fails to list the running guest
1274477112 M * fzylogic vc_rlimit_stat(RLIMIT_AS): Invalid argument
1274477124 M * fzylogic but the guest _is_ running as I can ssh into it
1274477130 M * Bertl that is kind-of fine
1274477140 M * Bertl i.e. the limit was removed on the kernel side
1274477162 M * fzylogic but it seems to be preventing vserver-stat from seeing the guest as running
1274477185 M * fzylogic or it's just a coincidence, but vserver-stat shows no running guests either way  :)
1274477187 M * Bertl we might need to either adjust util-vserver or provide some fake info
1274477214 M * fzylogic ok
1274477837 Q * richard_ Remote host closed the connection
1274478890 Q * petzsch Quit: Leaving.
1274479296 M * Bertl fzylogic: okay, I uploaded two patches, the first one is a cleanup, the second one will give you the memory limits back (although they will be 0)
1274479325 M * fzylogic is that preferable to fixing up util-vserver?
1274479512 M * Bertl no, but I guess it will take a while for daniel_hozac to switch to cgroup memory accounting
1274479520 M * fzylogic ok, that's fine
1274479522 M * fzylogic building now
1274479560 M * fzylogic annnd rebooting
1274479562 M * daniel_hozac do we have a vci version/bit for that?
1274479581 M * Bertl we can reserve one, or bump the version?
1274479586 M * Bertl what do you prefer?
1274479658 M * Bertl note that with memory cgroups disabled, that info is simply not available anymore
1274479662 M * daniel_hozac i think a version bump makes most sense.
1274479674 M * daniel_hozac well...
1274479691 M * daniel_hozac should we just start defaulting to using cgroups then?
1274479723 M * Bertl yes, I guess so, the question is if we should force it on
1274479744 M * Bertl IMHO there might be use cases where it simply isn't required
1274479764 M * Bertl (and I'd hate to hurt those by adding unneccesary overhead)
1274479785 M * daniel_hozac yeah
1274479802 M * Bertl so maybe we should default to on, but still allow to turn it off
1274479805 M * daniel_hozac i guess i'll just fall back on the old sum-up-processes.
1274479822 M * Bertl that's a good option too, IMHO
1274479830 J * FireEgl Proteus@2001:470:e056:1:223:54ff:fe89:b207
1274479847 M * Bertl okay, so we up the version then, and no vci bit, yes?
1274479868 M * daniel_hozac if we're not forcing it on, i'd prefer a VCI bit.
1274479884 M * daniel_hozac well, actually, i guess a version bump and a VCI bit.
1274479899 M * Bertl okay, and the vci goes on/off with the cgroup, yes?
1274479905 M * daniel_hozac the version bump to announce that the AS rlimit is gone, the VCI bit to say that memory cgroups are available.
1274479917 M * Bertl okay, got it
1274479946 M * fzylogic Bertl: patch seems to have fixed both the oops and the vserver-stat issue
1274480123 M * Bertl interesting ...
1274480208 M * Bertl daniel_hozac: http://vserver.13thfloor.at/ExperimentalT/delta-memcg-fix03.diff
1274480230 M * daniel_hozac looks good, thank you.
1274480236 M * Bertl you're welcome!
1274480389 M * fzylogic I even accessed /proc/meminfo from the guest to be sure  :)
1274480397 M * fzylogic stress-testing the OOM killer now
1274483915 Q * dowdle Remote host closed the connection
1274484855 M * fzylogic got the meminfo_proc_show bug to show up again
1274484996 M * fzylogic testing a couple of things to find the exact situation that causes it...
1274485148 M * fzylogic triggered by a vserver <guestname> stop
1274485161 M * fzylogic or (strangely) also if the guest's started via the init script
1274485167 M * fzylogic rather than manually
1274485607 M * Bertl could you upload the oops and run the addresses through addr2line?
1274485727 M * fzylogic http://paste.linux-vserver.org/15341
1274485772 M * fzylogic ffffffff810cc7e6?
1274485797 M * Bertl addr2line -e vmlinux ffffffff811353f8
1274485808 M * fzylogic meminfo.c:0
1274485820 M * Bertl then recompile with debug info
1274485833 M * Bertl (and try to trigger it again :)
1274486271 M * fzylogic oh, also I know it's a mainline bug but given how easily it can affect vserver users, it might be a good idea to include this patch in the vserver patch until it hits mainline:
1274486271 M * fzylogic http://userweb.kernel.org/~akpm/mmotm/broken-out/memcg-make-oom-killer-a-no-op-when-no-killable-task-can-be-found.patch
1274486277 M * fzylogic that's the oom killer infinite loop
1274486289 M * fzylogic I think at least one person on the mailing list hit the same bug
1274486364 M * fzylogic  /home/jeremy/kernels/linux-2.6.34/fs/proc/meminfo.c:42