1255651531 Q * saulus Remote host closed the connection
1255651659 J * saulus ~saulus@d047127.adsl.hansenet.de
1255652005 J * scientes ~scientes@174-21-201-131.tukw.qwest.net
1255652036 Q * scientes Remote host closed the connection
1255654227 J * uva_ bno@118-160-166-250.dynamic.hinet.net
1255654304 J * MooingLe1ur ~troy@shells195.pinchaser.com
1255654323 J * infowolfe_ ~infowolfe@c-76-105-242-186.hsd1.or.comcast.net
1255654329 Q * awk synthon.oftc.net larich.oftc.net
1255654329 Q * nkukard_ synthon.oftc.net larich.oftc.net
1255654329 Q * uva synthon.oftc.net larich.oftc.net
1255654329 Q * MooingLemur synthon.oftc.net larich.oftc.net
1255654329 Q * pinochle synthon.oftc.net larich.oftc.net
1255654329 Q * vorlon synthon.oftc.net larich.oftc.net
1255654329 Q * micah synthon.oftc.net larich.oftc.net
1255654329 Q * AndrewLee synthon.oftc.net larich.oftc.net
1255654329 Q * infowolfe synthon.oftc.net larich.oftc.net
1255654338 J * vorlon vorlon@minbar.dodds.net
1255654338 J * micah ~micah@micah.riseup.net
1255654364 J * pinochle ~pinochle@adsl-76-193-171-186.dsl.chcgil.sbcglobal.net
1255654372 J * nkukard_ ~nkukard@196.212.73.74
1255654378 J * AndrewLee ~andrew@u7.hlc.edu.tw
1255654478 J * awk ~phillip@gw1.security.web.za
1255659464 M * Bertl off to bed now ... have a good one everyone!
1255659468 N * Bertl Bertl_zZ
1255660935 Q * Pazzo Quit: Bye!
1255662038 Q * hparker Quit: Read error: 104 (Peer reset by connection)
1255662046 J * saulus_ ~saulus@c135210.adsl.hansenet.de
1255662453 Q * ksn Ping timeout: 480 seconds
1255662456 Q * saulus Ping timeout: 480 seconds
1255662458 N * saulus_ SauLus
1255663460 J * ksn ~kuben@82.128.69.23
1255663485 J * balbir_ ~balbir@122.172.6.48
1255664514 P * vorlon 
1255669222 J * Shinsaku ~Shinsaku@chello089076140236.chello.pl
1255669732 Q * balbir_ Quit: Ex-Chat
1255669748 J * balbir_ ~balbir@122.172.6.48
1255669996 J * sharkjaw ~gab@90.149.121.45
1255670646 M * awk Ok good night Bertl_zZ
1255671773 Q * Mr_Smoke Ping timeout: 480 seconds
1255671801 Q * ksn Ping timeout: 480 seconds
1255672340 J * Mr_Smoke ~smokey@layla.lecoyote.org
1255673208 J * Mr_Smoke_ ~smokey@layla.lecoyote.org
1255673220 Q * Mr_Smoke Read error: Connection reset by peer
1255674928 Q * Mr_Smoke_ Ping timeout: 480 seconds
1255675152 J * Mr_Smoke ~smokey@layla.lecoyote.org
1255675843 Q * derjohn_foo Ping timeout: 480 seconds
1255675907 Q * Shinsaku 
1255676145 N * Bertl_zZ Bertl
1255676149 M * Bertl morning folks!
1255676296 J * dholbach ~daniel@i59F72D8A.versanet.de
1255676467 M * awk morning Bertl_zZ
1255676488 M * arachnist moin
1255677803 J * kir ~kir@swsoft-msk-nat.sw.ru
1255678028 J * derjohn_mob ~aj@80.85.196.112
1255679358 Q * Mr_Smoke Read error: Connection reset by peer
1255679363 J * Mr_Smoke ~smokey@layla.lecoyote.org
1255680122 Q * harry Ping timeout: 480 seconds
1255680527 J * dholbach_ ~daniel@i59F777A0.versanet.de
1255680963 Q * dholbach Ping timeout: 480 seconds
1255680997 J * mrfree ~mrfree@host1-89-static.40-88-b.business.telecomitalia.it
1255681015 J * Shinsaku ~Shinsaku@chello089076140236.chello.pl
1255681120 N * dholbach_ dholbach
1255681265 J * ghislainocfs2 ~Ghislain@adsl2.aqueos.com
1255681595 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d
1255681788 J * thierryp ~thierry@zanzibar.inria.fr
1255682666 J * harry ~harry@d51A461B4.access.telenet.be
1255684415 J * gnuk ~F404ror@pla93-3-82-240-11-251.fbx.proxad.net
1255685773 Q * awk Ping timeout: 480 seconds
1255686546 J * todo HydraIRC@89.72.37.91
1255690564 M * Bertl off for now .. bbl
1255690568 N * Bertl Bertl_oO
1255691895 Q * Shinsaku Read error: Connection reset by peer
1255692043 J * Shinsaku ~Shinsaku@chello089076140236.chello.pl
1255692209 J * awk ~phillip@gw1.security.web.za
1255692275 Q * infowolfe_ Read error: Operation timed out
1255692655 J * infowolfe_ ~infowolfe@c-76-105-242-186.hsd1.or.comcast.net
1255693752 M * awk Can somebody tell me how this was possible.. right, I have a windows 2007 workstation i'm using (laptop) I unplug it from the network, I close the lid... I head off to a client, do some work come back... open my laptop plug into the network and i'm still ssh'ed to 3 servers?
1255693773 M * awk Am I missing something? does suspend do something that can hold the session open in weird and wonderful way?
1255693842 J * pmenier ~pmenier@ACaen-152-1-28-228.w83-115.abo.wanadoo.fr
1255693961 M * PowerKe actually, if you don't close the session it remains open.
1255693988 M * awk How would it still open with no network connectivity? well the network was unplugged for 3 hours?
1255693994 M * PowerKe as long as there's no traffic on the ssh session, the client or server can't see if there's a network problem (since there are no missing ACKs)
1255694002 M * awk ahhhh
1255694034 M * awk ok, thanks.. seeing a few weird things have been happening today
1255694035 M * PowerKe Usually it's things like stateful firewalls that get in between because they usually require some activity to keep the port open
1255694071 M * PowerKe also, if one of the servers would have sent some output (say you were tailing a logfile), then they'd notice the network is down
1255694211 J * geb ~geb@earth.gebura.eu.org
1255694308 J * doener_ ~doener@i59F5487B.versanet.de
1255694413 Q * doener Ping timeout: 480 seconds
1255694838 Q * brccc Ping timeout: 480 seconds
1255695155 M * awk ahh, i see. thanks for covering that up
1255696088 Q * geb Remote host closed the connection
1255696611 J * Pazzo ~ugelt@reserved-225136.rol.raiffeisen.net
1255697307 Q * sharkjaw Quit: Leaving
1255698029 J * hparker ~hparker@2001:470:1f0f:32c:290:96ff:fe50:40fa
1255698944 J * geb ~geb@173.4.82-79.rev.gaoland.net
1255699931 Q * xdr Ping timeout: 480 seconds
1255700725 M * ghislainocfs2 awk: ssh has also some keepalive parameter to detect and close idle connexions
1255700860 P * kir Leaving.
1255700937 M * geb -o TCPKeepAlive
1255700976 M * geb but i think it is for keep connexions opened , if you have a poor nat
1255700986 M * geb you can also use sysctl
1255701010 M * geb net.ipv4.tcp_keepalive_time
1255701615 J * brccc ~bruce@72.20.27.65
1255701643 M * thermoman hi brccc alias `I'm too lame to read BitchX.doc'
1255702037 Q * mrfree Quit: Leaving
1255702732 Q * nou Ping timeout: 480 seconds
1255702769 J * nou Chaton@2001:6f8:328:bbc:6666:6667::
1255702953 M * Mr_Smoke Hm
1255702969 M * Mr_Smoke French host, UK v6 broker, what gives ? Hello nou :)
1255703182 M * _Shiva_ and from looking.at the IP it seems to be clear that this host ist IRC specific :-)
1255703202 M * Mr_Smoke No sh*t :)
1255703219 M * Mr_Smoke The :6666:6667 part was quite a giveaway :p
1255703962 M * sid3windr =)
1255703970 M * sid3windr SIXIXIXEVEN
1255703981 M * sid3windr I wuv it
1255703992 M * sid3windr my both recursors have .53 in their respective /24s :>
1255704100 M * Mr_Smoke Hehe
1255704111 M * Mr_Smoke I tend to number them after the xid's
1255704121 M * Mr_Smoke And I numbered the xid's sequentially
1255704126 M * Mr_Smoke I'm funny like that :p
1255704194 M * sid3windr :]
1255704274 M * _Shiva_ who needs port numbers in a IPv6 world :-)
1255704567 Q * nou Ping timeout: 480 seconds
1255704739 J * nou Chaton@causse.larzac.fr.eu.org
1255705206 J * marl_scot ~marl_scot@84.92.193.225
1255705789 M * marl_scot hi, can anyone point me inthe rite direction to fix the following problem? (using vserver 2.2.0.6-test / 0.30.215-pre2697) after building a server with ubuntu hardy, did an upgrade to jaunty, then tried an upgrade to karmic, all woirked fine until trying to upgrade to karmic, when i get errors reported from package upstart saying 'telinit no such process'
1255705831 M * ghislainocfs2 is this package usefull in a guest ? if not just remove it and restart the process
1255705857 M * ghislainocfs2 if its playing with telinit i think this is not usefull in a guest
1255705925 M * marl_scot lol, didnt even think of that as it was required by 'ubuntu-minimal, forgot that the ubuntu-minimul is just a virtual package, and can be removed!
1255705965 M * marl_scot its somethe to do with logging, think its kernal logging and system logging, so supose it isnt realy required
1255705968 M * marl_scot thanks
1255706143 M * pmenier Hi all
1255706176 M * pmenier little problem with 2.6.31.4-vs2.3.0.36.19
1255706229 M * pmenier i successfully build the new kernel but when i type uname -a it returns the name of a running vserver not the hostname of the host ??
1255706295 M * ghislainocfs2 the klernel login part yes :)
1255706315 M * ghislainocfs2 pmenier: you have utils that are too old
1255706342 M * pmenier ahh .... ok here the details: http://paste.linux-vserver.org/13513
1255706346 M * ghislainocfs2 pmenier: humm let me check i had this issue with  old tool 
1255706390 M * pmenier i use util-vserver-0.30.215 but i've tried to build util-vserver-0.30.216-pre2849 but i get an error when i launch make
1255706417 M * ghislainocfs2 use  util-vserver-0.30.216-pre2849.tar.bz2 23-Sep-2009 16:26  711K  Patches
1255706426 M * ghislainocfs2 http://people.linux-vserver.org/~dhozac/t/uv-testing/
1255706426 Q * Shinsaku 
1255706435 M * ghislainocfs2 recent kernel patch require recent tools too
1255706452 M * ghislainocfs2 oh sorry
1255706464 M * ghislainocfs2 you tried lol, not readed to the end
1255706475 M * ghislainocfs2 which error do you get ?
1255706476 M * pmenier http://paste.linux-vserver.org/13514
1255706537 M * ghislainocfs2 hum don't know
1255706563 M * pmenier perhaps should i try another version of util-vserver ?
1255706568 M * ghislainocfs2 hope daniel_hozac will read you and help
1255706577 M * ghislainocfs2 yes you can try the few last ones
1255706600 M * ghislainocfs2 i had this bug some month ago so the 5 lasts one should do fine
1255706639 M * pmenier ok trying now pre2848
1255706724 M * pmenier same error
1255706749 M * pmenier but i see that: python/_libvserver.c:27:20: error: Python.h: No such file or directory
1255706752 M * daniel_hozac pmenier: do you have python-devel installed? (or your distro's equivalent)
1255706770 M * pmenier hello daniel i'm looking
1255706825 M * daniel_hozac i'll improve configure to actually run a build test for that.
1255706872 M * ghislainocfs2 python ? i was thinking everyone developped in ruby those days ? (just kidding lol)
1255706906 M * dholbach gonzo: around?
1255707008 M * pmenier ok build complete but now i can't stop running vserver: vcontext: pivot_root(): Invalid argument
1255707087 M * daniel_hozac pivot_root should only be executed when you're starting a guest.
1255707111 M * daniel_hozac or, well, if the context doesn't already have it setup.
1255707124 M * daniel_hozac that sounds like your kernel is broken though.
1255707168 M * pmenier humm ... i've killed process linked to the vserver ,restarting it and stopping it: it seems ok
1255707169 M * daniel_hozac oh, no. you'll just have to stop the guest manually.
1255707172 M * daniel_hozac yeah.
1255707298 M * pmenier in fact i didn't stop any running guest before making install of the new util-vserver.
1255707601 M * pmenier ok all is right now. Thanks to all
1255707832 J * imcsk8 ~ichavero@nat.ti.uach.mx
1255707898 M * derjohn_mob dholbach, ehlo :O !
1255707913 M * dholbach hey derjohn_mob!
1255707916 M * derjohn_mob dholbach, gonzo still debugging your guest ?
1255707930 M * derjohn_mob yes, the "mob"-variant right now :)
1255707932 M * dholbach derjohn_mob: I don't know - he told me to find a few things out about karmic, upstart and stuff
1255707946 M * dholbach and I was able to get some more information
1255707979 M * dholbach it'd be nice if we found the minimal configuration/setup changes and could document them both in the ubuntu release notes and on the vserver wiki
1255707989 M * dholbach so people won't be hit too hard by the changes
1255708047 M * derjohn_mob personally I dont believe in a probelm mit karmic generally
1255708064 M * derjohn_mob are there any hints about karmic as guest being problematic?
1255708067 M * dholbach derjohn_mob: which problem are you referring to?
1255708079 M * dholbach maybe we are not talking about the same thing :)
1255708091 M * derjohn_mob dholbach, dont know - gonzo told my after karmic upgrade there are segfaults in the guest
1255708106 M * dholbach oh, that one's fixed
1255708113 M * dholbach we can't really tell where it came from
1255708125 M * dholbach probably something in the upgrade of libc went wrong and everything segfaulted
1255708126 M * dholbach it was no fun
1255708140 M * dholbach the new problem is: https://bugs.launchpad.net/upstart/+bug/430224/
1255708157 M * dholbach which as I see it is going to hit every vserver-karmic guest
1255708196 M * dholbach (comment 5)
1255708271 M * derjohn_mob dholbach, maybe you should talk to an ubuntu-specialist?
1255708276 M * derjohn_mob :-)
1255708282 M * dholbach haha! :)
1255708298 M * dholbach I tried to bring vserver specialists and ubuntu/upstart specialists together
1255708299 M * derjohn_mob hm, the startup-scripts in karmic are not sysv-style anymore?
1255708306 M * dholbach some
1255708318 M * derjohn_mob maybe we must do initstyle=plain for that kind of guest
1255708329 M * derjohn_mob (or how was that style called ? IIRc it was plain ....)
1255708333 M * dholbach "I echoed "plain" into /etc/vservers/<vservername>/apps/init/style"
1255708338 M * dholbach is what gonzo did last night
1255708347 M * dholbach that doesn't get us all the way yet
1255708361 M * dholbach  <Bertl> a) how is upstart started from initramfs on a normal system?
1255708361 M * dholbach   b) how is upstart signalled to shutdown/reboot
1255708361 M * dholbach   c) how to specify a console, but that is maculation
1255708374 M * dholbach and this is the answers I got:
1255708374 M * dholbach a) it's the last line in the initramfs "init" script.  After mounting the root filesystem, moving over things like /dev and stuff, it execs /sbin/init on the real root
1255708374 M * dholbach  b) the shutdown tool uses d-bus to ask upstart to emit a runlevel 0 (or 6) event, everything happens based off what's configured off that in /etc/init
1255708374 M * dholbach  c) upstart only uses "the console", which is either the active VT, or what you specify on the kernel with console= (eg. console=ttyS0)
1255708380 M * derjohn_mob strange, because the init should work "within" the guest, encapsulated with it's own init.
1255708388 M * dholbach also:
1255708388 M * dholbach <cjwatson> dholbach: tail -n2 /usr/share/initramfs-tools/init
1255708388 M * dholbach  dholbach: how to specify a console: console= boot parameter as usual, I'd have thought
1255708388 M * dholbach (15:46:23) Daniel: daniel@miyazaki:~$ tail -n2 /usr/share/initramfs-tools/init
1255708388 M * dholbach exec run-init ${rootmnt} ${init} "$@" <${rootmnt}/dev/console >${rootmnt}/dev/console 2>&1
1255708388 M * dholbach panic "Could not execute run-init."
1255708390 M * dholbach daniel@miyazaki:~$ 
1255708615 M * derjohn_mob hm, I just ask myself how mounting in initstyle plain is realized... I mean if the guest doesnt as usualy not have capabilities to mount stuf
1255708618 M * derjohn_mob stuff
1255708648 M * dholbach there's a new mountall package that's involved somewhere in there
1255708652 M * dholbach but I honestly don't know
1255708778 M * derjohn_mob mountall package in ubuntu ?
1255708787 M * dholbach yep, in karmic
1255708805 M * derjohn_mob well I was referring to the vserver's capability
1255708811 M * dholbach ah ok
1255708816 M * dholbach sorry, that I don't know
1255708819 M * derjohn_mob usuall you cannot mount anything
1255708836 M * derjohn_mob for a test we could try what happens if we give him that permission 
1255710569 Q * dholbach Quit: Ex-Chat
1255711493 Q * derjohn_mob Ping timeout: 480 seconds
1255711568 Q * awk Ping timeout: 480 seconds
1255711748 Q * gnuk Quit: NoFeature
1255713129 Q * Pazzo Quit: Bye!
1255713235 N * Bertl_oO Bertl
1255713251 M * Bertl back now ...
1255713276 J * uva bno@118-160-166-250.dynamic.hinet.net
1255713383 J * click_ click@ti0127a340-0204.bb.online.no
1255713452 Q * infowolfe_ resistance.oftc.net oxygen.oftc.net
1255713452 Q * pinochle resistance.oftc.net oxygen.oftc.net
1255713452 Q * MooingLe1ur resistance.oftc.net oxygen.oftc.net
1255713452 Q * uva_ resistance.oftc.net oxygen.oftc.net
1255713452 Q * FireEgl resistance.oftc.net oxygen.oftc.net
1255713452 Q * nenolod resistance.oftc.net oxygen.oftc.net
1255713452 Q * Guest999 resistance.oftc.net oxygen.oftc.net
1255713452 Q * click resistance.oftc.net oxygen.oftc.net
1255713452 N * click_ click
1255713549 J * infowolfe_ ~infowolfe@c-76-105-242-186.hsd1.or.comcast.net
1255713549 J * pinochle ~pinochle@adsl-76-193-171-186.dsl.chcgil.sbcglobal.net
1255713549 J * MooingLe1ur ~troy@shells195.pinchaser.com
1255713549 J * uva_ bno@118-160-166-250.dynamic.hinet.net
1255713549 J * nenolod ~nenolod@petrie.dereferenced.org
1255713549 J * Guest999 ~Genghis@ph34r.my.d-n-s.org.uk
1255713551 Q * uva_ Max SendQ exceeded
1255713622 Q * thierryp Quit: ciao folks
1255714252 J * FireEgl FireEgl@173-16-9-10.client.mchsi.com
1255714532 M * Mr_Smoke Bertl: ever seen vserver (stable, 2.2.0.7) going kernel panic on boot, in VirtualBox ?
1255714628 M * Bertl nope, reproduceable?
1255714810 M * Mr_Smoke Well it might mean I b0rked my config
1255714818 M * Bertl do you have a kernel trace to snapshot and upload somewhere?
1255714833 M * Mr_Smoke but all I did was cp the .config from a working kernel to the vserver source and make the new kernel
1255714854 M * Mr_Smoke Nope, I'm new to virtualbox so I can't even scroll the console to see when it happens at all
1255714873 M * Mr_Smoke This is all part of the effort to recreate the PHP/SEGMEXEC Grsec bug
1255714882 M * Bertl but you see something on the console, yes?
1255714891 M * Mr_Smoke Yeah, I see the actual kernel panic
1255714916 M * Bertl then upload that .. and btw, kvm is a lot simpler and debug friendlier :)
1255714947 M * Bertl you can easily boot via serial console on kvm, and get a textual trace whenever something goes wrong
1255714986 M * Mr_Smoke Yeah, and what would be even easier would be a native linux boot :p
1255715007 M * Bertl not necessarily
1255715011 M * Mr_Smoke I have yet to try KVM though
1255715028 M * Bertl a virtual machine has its advantages on testing
1255715032 M * Mr_Smoke True.
1255715054 M * Mr_Smoke But before I bug you guys with my panic, I'll try to make sure it's not a PEBKAC
1255715092 M * Bertl most likely it is, so simply upload it and we're done :)
1255715113 M * Mr_Smoke Ok then, I will (sometime this weekend or maybe monday)
1255715170 M * Bertl so be it ...
1255715199 M * Mr_Smoke Thanks :)
1255715216 M * Mr_Smoke I don't feel too comfy with KVM yet (as in, I've never tried it)
1255716292 J * Keybuk ~scott@halo.netsplit.com
1255716301 M * Bertl welcome Keybuk!
1255716324 M * Keybuk hey
1255716340 M * Bertl didn't see that you were on the CC when I sent the email reply :)
1255716388 M * Keybuk so I guess it'd be interesting to find out where vserver and upstart aren't quite meeting in the middle
1255716410 M * Bertl yep, to make all those milions of ubuntu users happy :)
1255716437 M * Bertl first, you know how Linux-VServer works (in regard of 'startup')?
1255716441 M * Keybuk no, not at all
1255716456 M * Bertl okay, let me give you a quick intro then
1255716507 M * Bertl Linux-VServer is a so called OS Level Virtualization technology ... recently I prefer to see it as Userspace Isolation technology ... think chroot on steroids
1255716520 M * Keybuk right, a bit like FreeBSD jails
1255716549 M * Bertl exactly, so, everything hardware related happens on the host, the guest ist basically hardware independant
1255716580 M * Bertl thus, a properly configured guest (on most distros) only starts the services you actually want to be running
1255716598 M * Keybuk how much separation is there between guests? do they have different pid namespaces, for example?
1255716602 M * Keybuk do they have different networking namespaces?
1255716614 M * Bertl pid yes, networking is optional
1255716626 M * Keybuk you misunderstand me
1255716636 M * Keybuk if one guest listens on an AF_UNIX abstract address of, say, /foo/bar
1255716637 M * Bertl for performance and to simplify administration, we provide network isolation
1255716646 M * Keybuk can another guest listen on that address, or does it get address in use?
1255716658 M * Keybuk can guests connect to each other's AF_UNIX sockets
1255716663 M * Keybuk (ie. can you cross-connect D-Bus)
1255716674 M * Bertl if you configure the guests properly, yes
1255716678 M * Bertl by default, no
1255716679 M * Keybuk ok
1255716702 M * Keybuk that's obviously quite important for Upstart, because it listens on a "well known" AF_UNIX socket
1255716721 M * Keybuk you can't have multiple instances on the same machine unless they're in different namespaces
1255716722 M * Bertl that's fine, and I don't think that will cause any problems
1255716736 M * Keybuk and obviously each guest would have to be in a different pid namespace so Upstart can be pid 1 in each
1255716746 M * Bertl yes and no again
1255716769 M * Bertl a separate pid space is not required, as we allow on shared pid spaces the init virtualization
1255716790 M * Bertl i.e. for a specific init process, pid = 1 will be shown, regardless of the actual pid
1255716792 M * Keybuk ok
1255716803 M * Bertl otherwise sysv init won't work either
1255716807 M * Keybuk indeed
1255716815 M * Keybuk the reason Upstart cares is *only* to be compatible with sysvinit
1255716830 M * Keybuk if I had my way, I'd break the fingers of all those people who type "init 2" in the system and expect it to do something
1255716837 M * Bertl that's interesting, as it might allow to remove that virtualization 
1255716854 M * Bertl (or at least disable it for upstart)
1255716873 M * Bertl given there is an option to configure upstart for that, of course
1255716878 M * Keybuk there isn't yet
1255716887 M * Keybuk so, AIUI, you have two configuration options
1255716896 M * Keybuk in one, the guest doesn't get a real init daemon
1255716899 M * Keybuk in the other, the guest does
1255716900 M * Keybuk is that right>
1255716919 M * Bertl correct, and the second one can be broken down in with and without init pid=1
1255716958 M * Bertl the init-less setup is called 'sysv init style'
1255716981 M * Keybuk it's not really sysv init style
1255716984 M * Bertl where the init case (with pid=1) is called 'plain init stule'
1255716985 M * Keybuk because /etc/inittab won't work
1255716987 M * Bertl *style
1255716990 M * Keybuk (unless you emulate that)
1255717004 M * Bertl nope, but the sysv runlevel scripts will get executed
1255717023 M * Keybuk right, for any active init daemon (upstart, launchd, initng, etc.) you're going to need to use the plain init style ;-)
1255717035 M * Keybuk because starting services is done by the init daemon, rather than by shell scripts
1255717036 M * Bertl correct
1255717067 M * Keybuk so in your plain init style, you exec the init daemon with pid 1 ?
1255717116 M * Bertl by default, yes
1255717164 M * Keybuk ok
1255717175 M * Bertl I think the main issues upstart might have is that by default, a guest doesn't have a console and /proc/cmdline entry
1255717195 M * Keybuk that shouldn't matter
1255717202 M * Keybuk as long as it has /dev/console or /dev/null you'll be fine
1255717213 M * Bertl okay, /dev/null is there
1255717224 M * Keybuk Upstart doesn't read /proc/cmdline - it doesn't need to, the kernel passes it as init's own argv[]
1255717227 M * Keybuk (and it doesn't use it anyway(
1255717244 M * Keybuk so I guess the first problem you have is where do the error messages go? :p
1255717247 M * Bertl soo ... folks told me that upstart refuses to start any services, just sits around and waits ... any ideas what might cause that?
1255717252 M * Keybuk no, no ideas
1255717262 M * Keybuk first thing is to set up a vserver
1255717272 M * Keybuk give it a useful /dev/console so you can see the error messages
1255717277 M * Keybuk and exec init with --verbose in its arguments
1255717279 M * Keybuk and see what happens
1255717306 M * Keybuk almost certainly it'll tell you exactly what's wrong, and it's just been missed before because it's gone to /dev/null
1255717313 M * Bertl that should be rather trivial, I presume there is an url where I can debootstrap this particular ubuntu version?
1255717381 M * Keybuk usual URLs
1255717390 M * Keybuk debootstrap karmic /somewhere
1255717413 J * hijacker_ ~hijacker@87-126-142-51.btc-net.bg
1255717466 M * Bertl got one for me? I'm not an ubuntu person :)
1255717473 J * derjohn_mob ~aj@tmo-108-154.customers.d1-online.com
1255717493 M * daniel_hozac http://ftp.ubuntu.com/ubuntu should work, IIRC
1255717526 M * Keybuk http://archive.ubuntu.com/ubuntu/ ;-)
1255717560 M * Bertl okay, give me a second to update to latest util-vserver to rule out any issues there
1255717707 M * Bertl so, let's talk about upstart options ... are there any? how do they get passed to upstart?
1255717731 M * Bertl obviously --verbose is one :)
1255717763 M * Bertl daniel_hozac: any way to pass options to init (without wrapping it)?
1255717789 M * daniel_hozac apps/init/runlevel.start
1255717793 M * Keybuk that's about the main one :p
1255717807 M * Bertl good choice, IMHO :)
1255717838 M * Bertl daniel_hozac: I just put the binary path and options there on a line?
1255717840 M * Keybuk I can't think, offhand, of any other interesting command-line arguments for init
1255717847 M * daniel_hozac Bertl: just options.
1255717855 M * Bertl ah, got it, tx
1255718024 M * Bertl daniel_hozac: seems debootstrap got renamed again ...
1255718147 M * Bertl daniel_hozac: does that look wrong?  'vserver vs103 build -m debootstrap --context 103 --hostname vs103.test.org --interface eth0:192.168.1.103/24 -- -d karmic -m http://archive.ubuntu.com/ubuntu/ -- -arch amd64
1255718212 M * daniel_hozac looks good to me.
1255718231 M * Bertl it immediately stops after downloading debootstrap 1.0.20 (adjusted the url)
1255718264 M * daniel_hozac util-vserver installed from RPMs?
1255718290 M * Bertl yep, built from tar, extracted via mkuvrel.sh
1255718341 Q * hijacker_ Ping timeout: 480 seconds
1255718453 M * daniel_hozac hmm
1255718475 M * Bertl doesn't look like the new debootstrap feels like doing something
1255718509 M * daniel_hozac right...
1255718558 M * Bertl Keybuk: does ubuntu provide a debootstrap as deb package?
1255718644 M * daniel_hozac Bertl: ah, try two dashes for arch?
1255718650 M * daniel_hozac works better for me, anyway.
1255718652 M * Bertl *damn again :)
1255718682 M * Keybuk yes
1255718685 M * Bertl yep, works
1255718758 M * Bertl Keybuk: does upstart (or the scripts executed from it) do something 'special' to the console?
1255719003 J * hijacker_ ~hijacker@87-126-142-51.btc-net.bg
1255719115 M * Keybuk not really
1255719122 M * Keybuk it resets it, but in the same way that sysvinit did
1255719661 Q * hijacker_ Ping timeout: 480 seconds
1255719672 J * hijacker_ ~hijacker@87-126-142-51.btc-net.bg
1255719818 M * Bertl okay, that should be fine
1255720059 M * Bertl well, I tried with /dev/tty (as console) and /dev/ttyS0, in any case, /sbin/init is hanging around as reported, and doesn't log anything, the apps/init/runlevel.start contains --verbose, which seems not to be passed, daniel_hozac?
1255720120 M * Bertl Keybuk: but I seem to get coredumps :)
1255720140 M * Bertl /vservers/vs103/core.3326: ELF 64-bit LSB core file x86-64, version 1 (SYSV), SVR4-style, from '/sbin/init --verbose
1255720163 M * Bertl daniel_hozac: so full reverse, --verbose is passed :)
1255720169 M * daniel_hozac :)
1255720180 M * Bertl Keybuk: want to get the core dumps?
1255720244 Q * todo Read error: Connection reset by peer
1255720258 J * todo HydraIRC@89.72.37.91
1255720374 M * Keybuk sure, which arch i386 or amd64
1255720392 M * Bertl x86-64, so amd64 for ubuntu
1255720414 M * Bertl want one for i386? should be easy to install a 32bit guest
1255720415 M * Keybuk argh
1255720423 M * Keybuk oh, no
1255720425 M * Keybuk amd64 is good
1255720427 M * Keybuk that's what I'm on here
1255720435 M * Keybuk could you compress the dump and upload it somewhere I can grab it
1255720457 M * Bertl http://vserver.13thfloor.at/Stuff/UPSTART/
1255720461 M * Bertl not worth compressing
1255720765 Q * Keybuk Read error: Connection reset by peer
1255720771 J * Keybuk_ ~scott@halo.netsplit.com
1255720807 M * Keybuk_ crashed out there
1255720809 N * Keybuk_ Keybuk
1255721333 M * Bertl so I take it, you can recreate the issue somehow?
1255721391 Q * lnostdal Quit: lnostdal
1255721615 M * Keybuk no
1255721619 M * Keybuk unrelated
1255721693 M * Bertl k, when you need something tested, just let me know
1255721733 M * Keybuk well, did you get that core dump?
1255721753 J * [1]todo HydraIRC@89.72.37.91
1255721757 M * Bertl yep, and uploaded it, didn't you get the url?
1255721764 M * Bertl http://vserver.13thfloor.at/Stuff/UPSTART/
1255721774 M * Keybuk no, must have crashed about that point
1255721805 M * Bertl JFYI, the channel is real-time logged, if that happens again check out irc.13thfloor.at
1255721807 Q * todo Read error: Connection reset by peer
1255721807 N * [1]todo todo
1255721836 M * Keybuk Bertl: both 403
1255721880 M * Bertl secyeah, forgot, coredumps are secret :)
1255721885 M * Bertl try again please
1255722042 M * Keybuk that's odd
1255722104 M * Keybuk Bertl: which upstart package did you use?
1255722106 M * Keybuk what was the version?
1255722135 M * Bertl the one which came with karmic, I'd say?
1255722145 M * Keybuk ok -10?
1255722168 M * Keybuk it looks like /dev/console isn't right
1255722203 M * Bertl how so?
1255722265 M * Keybuk it failed to open
1255722274 M * Bertl 0.6.3-10 btw
1255722383 M * Bertl well, it's obviously there and I can write to it, stuff shows up on the serial console
1255722397 M * Keybuk $8 = {filename = 0x7f03c0e214c3 "system.c", line = 110, 
1255722397 M * Keybuk   function = 0x7f03c0e214e0 "system_setup_console", number = 6, 
1255722397 M * Keybuk   message = 0x7f03c2c82168 "No such device or address"}
1255722424 M * Keybuk ENXIO when it tried to open /dev/console
1255722458 M * Bertl okay, let's try again, maybe I copied the wrong device ...
1255722479 M * Keybuk I don't think it's the device node, are you sure it's hooked up right inside openvz?
1255722494 M * Bertl not using openvz here :)
1255722541 M * Bertl yep, my fault, was /dev/ttyS0 instead of ttyS1
1255722565 M * Bertl but I consider it a bug that upstart coredumps when not able to open the console
1255722576 M * Keybuk it didn't coredump
1255722582 M * Keybuk well, the child did
1255722584 M * Bertl so who did?
1255722586 M * Keybuk ie. the thing it was trying to run
1255722597 M * Keybuk pid 1 didn't coredump, pid 3326 did :p
1255722629 M * Bertl coredumps use the global pid, so it probably was upstart
1255722649 M * Bertl anyway, it spews out a bunch of messages
1255722660 M * Bertl but in the end, only one process is running
1255722663 M * Keybuk nah, it's definitely the child
1255722669 M * Keybuk ok, what are the "bunch of messages"
1255722713 M * Bertl http://paste.linux-vserver.org/13517
1255722748 M * Keybuk #
1255722748 M * Keybuk #
1255722748 M * Keybuk libudev: udev_monitor_enable_receiving: bind failed: Operation not permitted  
1255722748 M * Keybuk #
1255722748 M * Keybuk mountall:mountall.c:2898: Assertion failed in main: udev_monitor_enable_receiving (udev_monitor) == 0  
1255722757 M * Bertl hardware related stuff, like hwclock should not be started inside a guest, but they will not do any harm ...
1255722759 M * Keybuk the "mountall" program couldn't connect to udev
1255722772 M * Bertl okay, so?
1255722783 M * Bertl udev should not be run either inside a guest
1255722783 M * Keybuk so it can't run
1255722792 M * Keybuk ah, no, Ubuntu requires udev
1255722795 M * Keybuk all distributions do now
1255722797 M * Keybuk it's replaced HAL
1255722804 M * Bertl no hardware to handle, nothing to manage there
1255722808 M * Keybuk udev doesn't just do hardware
1255722812 M * Keybuk it's the kernel/userspace interface
1255722817 M * Keybuk though, that's not actually udevd itself
1255722821 M * Bertl not really
1255722826 M * Keybuk that's the kernel uevent netlink socket
1255722856 M * Bertl so no uevents from kernel side inside a guest either
1255722859 M * Keybuk right
1255722863 M * Keybuk that's a bug in vserver
1255722872 M * Bertl no, trust me, it isn't :)
1255722875 M * Keybuk at this point, you're not really providing a Linux guest anymore
1255722884 Q * bonbons Quit: Leaving
1255722894 M * daniel_hozac it's got no hardware.
1255722895 M * Bertl look, we are not doing hardware virtualization
1255722909 M * Keybuk so?  you still have to have a bridge between the kernel and the userspace
1255722912 M * Keybuk udev isn't just used for hardware
1255722923 M * daniel_hozac those are called syscalls.
1255722928 M * Keybuk not true
1255722934 M * Keybuk /proc isn't a syscall, I bet you mount that in the guest
1255722944 M * Bertl yes, we do, so?
1255722957 M * Keybuk so you need to provide the uevent netlink socket too
1255722961 M * Keybuk it's part of the Linux standard interface
1255722966 M * Keybuk along with /sys
1255722977 M * daniel_hozac show us a uevent that makes sense in a guest.
1255723010 M * Keybuk daniel_hozac: any for devices in /dev
1255723015 M * Keybuk unless you don't have a /dev in the guest :p
1255723020 M * daniel_hozac yeah, we don't.
1255723034 M * Keybuk but you have ttys there, Bertl said you do above
1255723035 M * Bertl there is a fixed set of devices in /dev and new ones cannot be created
1255723101 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d
1255723122 M * daniel_hozac things like /dev/tty, /dev/null, /dev/zero, etc. don't really generate events, AFAIK.
1255723126 M * Keybuk yes they do
1255723146 M * Bertl http://paste.linux-vserver.org/13518
1255723205 M * Bertl the console is normally not part of a guest
1255723237 M * Keybuk right, you do need a /dev/console
1255723243 M * Keybuk at least, if any file in /etc/init needs it
1255723249 M * Bertl that's why I added it for you :)
1255723263 M * Keybuk strictly speaking you probably don't need all this udev stuff
1255723272 M * Keybuk but then you'll need other things in their place to do things like mount the disk?
1255723274 M * Bertl now we are getting somewhere
1255723278 M * Keybuk (maybe your disk is already mounted)
1255723284 M * Bertl nope, everything is mounted
1255723293 M * Bertl no mounting is allowed (by default) either
1255723308 M * Keybuk is it acceptable for vserver to sed some files in the default install? :P
1255723313 M * Keybuk before starting the guest
1255723321 M * Keybuk because you could
1255723323 M * Bertl sure, we do a guest cleanup after install
1255723326 M * Keybuk  - not include udev, mountall, etc.
1255723338 M * Keybuk  - sed /etc/init/rc-sysinit.conf, change the "start on filesystem" line to "start on startup"
1255723355 M * Keybuk  - remove /etc/init/hwclock.conf
1255723378 M * Keybuk actually
1255723379 M * Keybuk instad
1255723382 M * daniel_hozac removing is the only way to disable it?
1255723384 M * Keybuk don't sed the file
1255723391 M * Keybuk create an /etc/init/vserver.conf
1255723393 M * Keybuk containing:
1255723395 M * Keybuk   start on startup
1255723397 M * Keybuk   script
1255723405 M * Keybuk     initctl emit virtual-filesystems
1255723411 M * Keybuk     initctl emit local-filesystems
1255723415 M * Keybuk     initctl emit remote-filesystems
1255723418 M * Keybuk     initctl emit filesystem
1255723419 M * Keybuk   end script
1255723428 M * Keybuk that requires less hackery
1255723434 M * Keybuk and then don't include udev, mountall, etc.
1255723571 M * Bertl include means?
1255723581 M * Bertl I remove them or what?
1255723937 J * ViRUS ~mp@p579B4F7E.dip.t-dialin.net
1255724405 M * Bertl Keybuk: still around?
1255724997 J * Pazzo ~ugelt@sadsl-246059.rol.raiffeisen.net
1255725020 Q * Pazzo 
1255725344 M * Keybuk Bertl: yeah remove the packages
1255725356 M * Keybuk or just the /etc/init/*.conf files for them
1255725362 M * Bertl okay
1255725383 J * steffan steffan@steffan.netrep.oftc.net
1255725409 Q * hijacker_ Remote host closed the connection
1255725492 P * steffan 
1255725687 M * Bertl daniel_hozac: okay, after reducing them to this: http://paste.linux-vserver.org/13519 the guest seems to start fine
1255725837 M * daniel_hozac Bertl: removing the files won't work though, as package managers will just put them back. i guess commenting them out if there's no way to just disable it is the way to go...
1255725856 M * Bertl Keybuk: any suggestions?
1255725877 M * Bertl Keybuk: maybe add a 'disable' option or so?
1255725934 M * Keybuk Bertl: dpkg will not replace a removed conffile
1255725962 M * daniel_hozac but rpm will.
1255725975 M * Keybuk Fedora doesn't use Upstart in this way yet
1255726098 M * Bertl nevertheless, I consider it at least a missing feature if you cannot 'disable' a 'runlevel' service
1255726109 M * Keybuk it is a missing feature
1255726162 M * Bertl okay, so let's comment them out for now, and when the feature is added, we switch to disabling them, sounds good?
1255726268 M * Keybuk yup
1255726290 M * Bertl good, then please drop us a note when you added that feature ...
1255726388 M * Bertl thanks for taking the time, feel free to hang around
1255726750 Q * Keybuk Read error: No route to host
1255728439 J * uva_ bno@118-168-238-45.dynamic.hinet.net
1255728642 J * blues_ ~blues@aga136.neoplus.adsl.tpnet.pl
1255728652 Q * uva Read error: Connection reset by peer
1255728763 Q * blues Ping timeout: 480 seconds
1255729330 Q * derjohn_mob Ping timeout: 480 seconds
1255731209 Q * uva_ synthon.oftc.net oxygen.oftc.net
1255731209 Q * Guest999 synthon.oftc.net oxygen.oftc.net
1255731209 Q * MooingLe1ur synthon.oftc.net oxygen.oftc.net
1255731209 Q * pinochle synthon.oftc.net oxygen.oftc.net
1255731209 Q * nenolod synthon.oftc.net oxygen.oftc.net
1255731209 Q * infowolfe_ synthon.oftc.net oxygen.oftc.net
1255731213 J * uva_ bno@118-168-238-45.dynamic.hinet.net
1255731213 J * infowolfe_ ~infowolfe@c-76-105-242-186.hsd1.or.comcast.net
1255731213 J * pinochle ~pinochle@adsl-76-193-171-186.dsl.chcgil.sbcglobal.net
1255731213 J * MooingLe1ur ~troy@shells195.pinchaser.com
1255731213 J * nenolod ~nenolod@petrie.dereferenced.org
1255731213 J * Guest999 ~Genghis@ph34r.my.d-n-s.org.uk
1255732298 Q * imcsk8 Quit: Leaving
1255732701 Q * todo Read error: Connection reset by peer
1255733889 J * derjohn_mob ~aj@e180193098.adsl.alicedsl.de
1255737078 Q * derjohn_mob Ping timeout: 480 seconds