1181952050 M * mugwump apt-get install manpages-dev
1181954198 J * dreamind apwdsl@p548AACFC.dip0.t-ipconnect.de
1181954199 Q * dreamind 
1181954541 M * coderanger daniel_hozac: Does get_nx_info do anything useful?
1181954572 J * DoberMann_ ~james@AToulouse-156-1-158-138.w90-38.abo.wanadoo.fr
1181954679 Q * DoberMann[PullA] Ping timeout: 480 seconds
1181954945 A * coderanger offlines for the weekend, catch y'all later
1181955126 J * click click@ti511110a080-0449.bb.online.no
1181955302 N * Bertl_oO Bertl
1181955307 M * Bertl back now ..
1181955424 M * Bertl daniel_hozac: just compiled 0.30.213 with minimal modifications (i.e. disabled doc and changed the required stuff to match mandrake) and installing the packages gives me: /usr/sbin/setattr: No such file or directory (which is not that unusual, as it is /sbin/setattr) shouldn't that be detected at build time?
1181955429 N * DoberMann_ DoberMann
1181955429 Q * coderanger Ping timeout: 480 seconds
1181955531 N * DoberMann DoberMann[ZZZzzz]
1181955567 M * Bertl daniel_hozac: please ignore that, this seems to be from the uninstall of the previous packages
1181961021 J * ktwilight ~ktwilight@235.114-66-87.adsl-dyn.isp.belgacom.be
1181961032 M * Bertl welcome ktwilight!
1181961404 Q * ktwilight_ Ping timeout: 480 seconds
1181962889 J * n9ab ~king@168.187.117.237
1181962897 M * Bertl welcome n9ab!
1181963028 Q * n9ab 
1181963159 Q * brcc_ Ping timeout: 480 seconds
1181963561 Q * infowolfe_ Ping timeout: 480 seconds
1181965133 J * infowolfe ~infowolfe@c-71-195-248-179.hsd1.ut.comcast.net
1181965326 J * brcc_ bruce@i.am.someasshole.com
1181965327 Q * infowolfe Read error: Connection reset by peer
1181966959 J * infowolfe ~infowolfe@c-71-195-248-179.hsd1.ma.comcast.net
1181969692 J * infowolfe_ ~infowolfe@c-71-195-248-179.hsd1.ut.comcast.net
1181969961 Q * infowolfe Ping timeout: 480 seconds
1181970515 M * Bertl okay, off to bed now ... have a good one everyone!
1181970534 N * Bertl Bertl_zZ
1181971456 Q * meandtheshel1 Quit: Leaving.
1181972705 J * Aiken_ ~james@ppp121-45-250-169.lns2.bne4.internode.on.net
1181973024 Q * Aiken Ping timeout: 480 seconds
1181973543 J * HeinMueck ~Miranda@dslb-088-064-029-026.pools.arcor-ip.net
1181973966 J * _jthm_ ~infowolfe@c-71-195-248-179.hsd1.ut.comcast.net
1181974384 Q * infowolfe_ Ping timeout: 480 seconds
1181974412 J * infowolfe ~infowolfe@c-71-195-248-179.hsd1.ut.comcast.net
1181974774 Q * _jthm_ Ping timeout: 480 seconds
1181974777 Q * Aiken_ Quit: Leaving
1181975040 J * infowolfe_ ~infowolfe@c-71-195-248-179.hsd1.ut.comcast.net
1181975319 J * dna ~naucki@147-199-dsl.kielnet.net
1181975434 Q * infowolfe Ping timeout: 480 seconds
1181975556 Q * infowolfe_ Ping timeout: 480 seconds
1181976786 J * ruskie_ ruskie@goatse.co.uk
1181976834 J * _Radiance f229ae6f98@halt.1984world.eu
1181976843 J * arachnis1 arachnist@088156185052.who.vectranet.pl
1181976848 J * _[PUPPETS]Gonzo gonzo@langweiligneutral.deswahnsinns.de
1181976860 J * pyquila gerbens@82.94.222.35
1181976867 J * AndrewLe1 ~andrew@flat.iis.sinica.edu.tw
1181976868 J * xe ex@valis.net.pl
1181976877 J * weasel_ weasel@asteria.debian.or.at
1181976879 J * eyck__ eyck@195.242.124.71
1181976882 J * trippeh_ atomt@uff.ugh.no
1181976887 J * phedny_ ~mark@86.83.129.67
1181976887 J * virtuoso_ ~s0t0na@80.253.205.251
1181976896 J * Bertl_zZ_ herbert@IRC.13thfloor.at
1181976899 Q * HeinMueck charon.oftc.net galapagos.oftc.net
1181976899 Q * pusling charon.oftc.net galapagos.oftc.net
1181976899 Q * dlezcano1 charon.oftc.net galapagos.oftc.net
1181976899 Q * weasel charon.oftc.net galapagos.oftc.net
1181976899 Q * ruskie charon.oftc.net galapagos.oftc.net
1181976899 Q * phedny charon.oftc.net galapagos.oftc.net
1181976899 Q * arachnist charon.oftc.net galapagos.oftc.net
1181976899 Q * eyck_ charon.oftc.net galapagos.oftc.net
1181976899 Q * AndrewLee charon.oftc.net galapagos.oftc.net
1181976899 Q * daniel_hozac charon.oftc.net galapagos.oftc.net
1181976899 Q * virtuoso charon.oftc.net galapagos.oftc.net
1181976899 Q * morfoh charon.oftc.net galapagos.oftc.net
1181976899 Q * derjohn charon.oftc.net galapagos.oftc.net
1181976899 Q * cehteh charon.oftc.net galapagos.oftc.net
1181976899 Q * [PUPPETS]Gonzo charon.oftc.net galapagos.oftc.net
1181976899 Q * ex charon.oftc.net galapagos.oftc.net
1181976899 Q * trippeh charon.oftc.net galapagos.oftc.net
1181976899 Q * Bertl_zZ charon.oftc.net galapagos.oftc.net
1181976899 Q * Radiance charon.oftc.net galapagos.oftc.net
1181976899 Q * pyquila_ charon.oftc.net galapagos.oftc.net
1181976899 N * arachnis1 arachnist
1181976899 N * xe ex
1181976906 J * daniel_hozac ~daniel@c-051472d5.08-230-73746f22.cust.bredbandsbolaget.se
1181976913 J * morfoh ~morfoh@kilo105.server4you.de
1181976913 N * weasel_ Guest2033
1181976915 N * Guest2033 weasel
1181976915 N * ruskie_ ruskie
1181976915 J * pusling pusling@88.212.70.38
1181977498 J * infowolfe ~infowolfe@c-71-195-248-179.hsd1.ma.comcast.net
1181977498 Q * infowolfe 
1181977523 J * derjohn ~derjohn@80.69.41.3
1181977534 J * dlezcano1 ~dlezcano1@AToulouse-252-1-110-148.w86-217.abo.wanadoo.fr
1181977567 J * cehteh ~ct@pipapo.org
1181979437 J * Aiken ~james@ppp121-45-250-169.lns2.bne4.internode.on.net
1181980726 J * meandtheshell ~markus@85-124-232-67.work.xdsl-line.inode.at
1181981529 J * HeinMueck ~Miranda@dslb-088-064-029-026.pools.arcor-ip.net
1181982370 Q * FireEgl Quit: Bye...
1181983001 J * FireEgl FireEgl@2001:5c0:84dc:1:ff:39ff:fed1:1f2d
1181983210 J * ktwilight_ ~ktwilight@121.113-66-87.adsl-dyn.isp.belgacom.be
1181983210 Q * ktwilight Read error: Connection reset by peer
1181983345 M * micah Bertl_zZ_: sorry was afk (travelling), but maybe I can now?
1181983419 J * ktwilight ~ktwilight@232.214-66-87.adsl-static.isp.belgacom.be
1181983801 Q * ktwilight_ Ping timeout: 480 seconds
1181985102 Q * meandtheshell Quit: Leaving.
1181985229 Q * eSa| Quit: KVIrc 3.2.4 Anomalies http://www.kvirc.net/
1181985544 J * infowolfe ~infowolfe@c-71-195-248-179.hsd1.ut.comcast.net
1181985547 Q * infowolfe 
1181985847 J * bulkBoy ~eSa|@ip-87-238-2-45.adsl.cheapnet.it
1181985873 J * daniele_ ~eSa|@ip-87-238-2-45.adsl.cheapnet.it
1181985875 Q * daniele_ 
1181986916 N * DoberMann[ZZZzzz] DoberMann
1181987001 M * daniel_hozac micah: i think it was about having recent Debian packages for the various distributions on ftp.linux-vserver.org, or at least linked from the wiki.
1181987054 M * arachnist what other distros? debian itself, adamantix, ubuntu and...?
1181987075 M * daniel_hozac stable and testing.
1181987084 M * arachnist ah
1181987092 M * waldi arachnist: no real difference
1181987134 M * daniel_hozac except Ubuntu's broken ;)
1181987147 M * waldi ubuntu is always broken
1181987187 M * daniel_hozac yeah, looks like...
1181987791 Q * ||Cobra|| Ping timeout: 480 seconds
1181987806 J * ensc ~irc-ensc@p54B4EBAB.dip.t-dialin.net
1181988087 J * tokkee tokkee@casella.verplant.org
1181988289 Q * FireEgl Quit: Bye...
1181988571 J * ema ~ema@rtfm.galliera.it
1181988621 Q * daniel_hozac Ping timeout: 480 seconds
1181988843 N * phedny_ phedny
1181989036 J * daniel_hozac ~daniel@c-051472d5.08-230-73746f22.cust.bredbandsbolaget.se
1181991748 J * FireEgl FireEgl@Atlantica.Inet6.Info
1181992110 Q * HeinMueck Quit: Aah!
1181992708 J * HeinMueck ~Miranda@dslb-088-064-029-026.pools.arcor-ip.net
1181993439 Q * meetra Ping timeout: 480 seconds
1181993654 J * meetra ~xtreme@bl6-157-8.dsl.telepac.pt
1181993991 Q * sladen Ping timeout: 480 seconds
1181995302 Q * HeinMueck Quit: Aah!
1181996740 J * meetra45 ~xtreme@bl6-154-162.dsl.telepac.pt
1181996935 Q * meetra Ping timeout: 480 seconds
1181997005 N * meetra45 meetra
1181997266 Q * brcc_ Ping timeout: 480 seconds
1181997367 J * meetra45 ~xtreme@bl6-154-162.dsl.telepac.pt
1181997415 Q * meetra45 
1181997606 N * AndrewLe1 AndrewLee
1181997775 Q * meetra Ping timeout: 480 seconds
1181997990 Q * j0lt Ping timeout: 480 seconds
1181998055 J * brcc_ bruce@i.am.someasshole.com
1181998272 Q * Aiken Quit: Leaving
1181998292 J * infowolfe ~infowolfe@c-24-10-234-129.hsd1.ut.comcast.net
1181998742 N * DoberMann DoberMann[PullA]
1181998813 Q * ema Quit: leaving
1182000490 Q * infowolfe Quit: Leaving
1182000538 J * ||Cobra|| ~cob@pc-csa01.science.uva.nl
1182002093 J * infowolfe ~infowolfe@c-24-10-234-129.hsd1.ut.comcast.net
1182002130 Q * ensc Ping timeout: 480 seconds
1182002420 J * infowolfe_ ~infowolfe@c-24-10-234-129.hsd1.ut.comcast.net
1182002467 M * micah daniel_hozac: thats a good idea, I used to keep those listed on the old site
1182002476 M * micah daniel_hozac: but have not recently
1182002577 M * daniel_hozac micah: will you be uploading 0.30.213 to backports.org?
1182002754 Q * infowolfe Ping timeout: 480 seconds
1182002811 M * micah daniel_hozac: yes, in fact I hope to do it in the next few days
1182002816 M * daniel_hozac okay, great!
1182002867 N * infowolfe_ infowolfe
1182002880 M * micah if VSERVER_REMAP_SADDR is not on you can still configure a 'lo' device and set it to 127.0.0.1 for a guest, right?
1182002900 M * daniel_hozac what is it that you want to accomplish?
1182002924 M * micah i was just responding to a bug report from someone who was complaining about 127.0.0.1 not being accessible
1182002942 M * daniel_hozac in what sense? TCP/UDP should work fine...
1182002967 M * daniel_hozac ping is about the only thing that doesn't work.
1182002985 M * daniel_hozac (regardless of kernel configuration)
1182003011 M * micah yea, I'm trying to figure out what this person is wanting to accomplish
1182003058 M * micah (or why they reported the bug to vserver-debiantools for that matter)
1182003061 M * micah if you want to see: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=426797
1182003085 M * daniel_hozac so when is vserver-debiantools getting retired? :)
1182003096 M * micah i was hoping to chase that some more this week too
1182003112 M * micah it seems you made some good progress with Ola in that bug report, but it went stale
1182003344 M * daniel_hozac yeah, i really hope he looks at integrating tha last few features into util-vserver instead...
1182004241 J * meetra ~xtreme@82.155.157.60
1182005010 J * oauto ~micah@micah.riseup.net
1182005011 Q * oauto 
1182007282 N * DoberMann[PullA] DoberMann
1182008317 J * HeinMueck ~Miranda@dslb-088-064-029-026.pools.arcor-ip.net
1182008814 M * HeinMueck when I get "vcontext: vc_create_context(): File exists" - which file could that be?
1182008943 J * onox ~onox@kalfjeslab.demon.nl
1182009184 M * daniel_hozac it means the context you've chosen already exists.
1182009184 Q * slacker403 Read error: Connection reset by peer
1182009630 M * HeinMueck An where? I'm confused :) What I did is I created /etc/vserver/balu/context filled with the value 4 and set the xid of the vserver root to 4 (-R) ...
1182009645 M * HeinMueck When I try to start it I get the message.
1182009662 M * HeinMueck I must be doing something stupid :)
1182009666 M * daniel_hozac which means that there's already a guest running with xid 4.
1182009824 J * ensc ~irc-ensc@p54B4DD3A.dip.t-dialin.net
1182009924 J * HeinMueck41 ~Miranda@dslb-088-065-247-184.pools.arcor-ip.net
1182009981 M * HeinMueck41 could it be there are artefacts in the system from earlier tries?
1182010042 M * daniel_hozac possibly.
1182010042 M * HeinMueck41 marvin:~# vserver-stat
1182010043 M * HeinMueck41 CTX   PROC    VSZ    RSS  userTIME   sysTIME    UPTIME NAME
1182010043 M * HeinMueck41 0       43  63.1M  20.5M   0m41s32   0m55s10   1d02h49 root server
1182010043 M * HeinMueck41 49183    3  33.5M   4.5M   0m00s14   0m00s20   3h34m23 arch
1182010043 M * HeinMueck41 marvin:~#
1182010061 M * HeinMueck41 ah, okay
1182010092 M * HeinMueck41 is there a way to find them? (just interest :)
1182010115 M * daniel_hozac to find what? existing contexts?
1182010126 M * daniel_hozac a recent vserver-stat would do, otherwise ls -l /proc/virtual.
1182010152 M * HeinMueck41 k, found a lot of old entries in /proc/virtual, also a file named 4 is there ...
1182010175 Q * HeinMueck Ping timeout: 480 seconds
1182010263 M * onox -+
1182010293 M * HeinMueck41 there seems to be a general problem with start/stop on rc and the missing INIT environment vars.
1182010330 M * HeinMueck41 Even a new etch guest on an etch host has probs stopping because of a missing PREVLEVEL env
1182010340 M * daniel_hozac hmm?
1182010353 M * onox oops, that's my keyboard
1182010366 M * daniel_hozac i've had no problems with my etch guests.
1182010387 M * HeinMueck41 let me see ...
1182010467 M * HeinMueck41 I have created a new guest from scratch with the debian tools
1182010475 M * HeinMueck41 when I stop it I get :
1182010476 M * HeinMueck41 cat: /proc/cmdline: No such file or directory
1182010476 M * HeinMueck41 mount: permission denied
1182010476 M * HeinMueck41 Warning: Executing wildcard deletion to stay compatible with old scripts.
1182010476 M * HeinMueck41          Explicitly specify the prefix length (78.47.248.162/32) to avoid this warning.
1182010476 M * HeinMueck41          This special behaviour is likely to disappear in further releases,
1182010478 M * HeinMueck41          fix your scripts!
1182010542 M * daniel_hozac that warning is because you didn't specify a prefix or a netmask in the configuration.
1182010568 M * daniel_hozac the mount and /proc/cmdline errors look like older utils again.
1182010797 M * HeinMueck41 mount seems to be a S60umountroot in rc6.d :)
1182010950 M * daniel_hozac ... which is cleaned up by recent utils.
1182011082 Q * dna Quit: Verlassend
1182011361 J * dna ~naucki@147-199-dsl.kielnet.net
1182011408 M * HeinMueck41 Is there a way to tell the kernel to remove a context without rebooting?
1182011423 M * HeinMueck41 It tried to start the server once, no errors, but also no server
1182011450 M * HeinMueck41 now I have the ctx still in /proc/virtual and get errors again
1182011693 M * HeinMueck41 I downloaded a suse guest image from lylix.net and it somehow has the same behavoir than the one I copied yesterday
1182011731 M * HeinMueck41 at the top of the rc srcipt they have test "$PREVLEVEL" = "$RUNLEVEL" && exit 0
1182011800 Q * daniel_hozac Read error: Connection reset by peer
1182012047 M * HeinMueck41 at the top the suse rc srcipt has the line : test "$PREVLEVEL" = "$RUNLEVEL" && exit 0
1182012066 M * HeinMueck41 which means exit always, because both vars are missing
1182012089 M * HeinMueck41 they also try to use $CONSOLE 
1182012302 J * daniel_hozac ~daniel@c-051472d5.08-230-73746f22.cust.bredbandsbolaget.se
1182012680 M * daniel_hozac HeinMueck41: a context will go away as soon as there's nothing keeping it, like a process, file handle, whatever.
1182012696 M * daniel_hozac (unless you made it persistent)
1182012790 N * Bertl_zZ_ Bertl
1182012798 M * Bertl morning folks!
1182012816 M * HeinMueck41 would you recommend always setting an explicit context for a vserver? 
1182012851 M * Bertl yes, and that will be done (with recent tools) anyways
1182012887 M * HeinMueck41 okay
1182012922 J * infowolfe_ ~infowolfe@c-24-10-234-129.hsd1.ut.comcast.net
1182012937 M * Bertl daniel_hozac: could you have a look at http://vserver.13thfloor.at/Experimental/delta-nsproxy-fix04.diff and comment?
1182012955 M * Bertl (only relevant for 2.6.22-rc3+
1182013105 M * daniel_hozac won't the removal of that get_nsproxy ripple through to other places than vx_migrate_task?
1182013123 M * daniel_hozac or is that the intention?
1182013165 M * Bertl I'm not 100% sure about that one, but IMHO that is a bugfix which might apply to older kernels too
1182013181 M * Bertl the clone gives us a proxy with refcount 1
1182013215 M * Bertl the various copy/mix versions do not put that at any time
1182013239 M * Bertl the comment above __vs_merge_nsproxy() states that it holds a reference
1182013272 M * Bertl so IMHO that is exactly what it does right now ... but please double check
1182013342 M * daniel_hozac the comment states that it _will_ hold a reference, no?
1182013344 Q * infowolfe Ping timeout: 480 seconds
1182013362 M * Bertl yep, and that is true, as we do not put the count=1 from the clone
1182013452 M * Bertl vs_mix_proxy -> copy_nsproxy -> clone_nsproxy
1182013475 M * Bertl clone_sproxy: atomic_set(&ns->count, 1)
1182013491 M * Bertl AFAIKT, there is no put on the way back up
1182013518 M * Bertl so we end up with a new proxy, with count=1, i.e. with a reference held
1182013531 M * daniel_hozac but vs_mix_proxy isn't called for that one.
1182013546 M * daniel_hozac i mean, if vs_mix_proxy is called, we're in a different branch of that function.
1182013568 M * Bertl ah, right, okay, then that one is wrong ...
1182013581 M * Bertl I should add a comment there :)
1182013587 M * daniel_hozac or at least else { ;)
1182013600 M * Bertl yeah, unfortunately that makes gcc complain, IIRC
1182013621 M * Bertl but I can do the {} for the if
1182013667 M * daniel_hozac hmm, why would gcc complain?
1182013689 M * Bertl some versions do not know that the function always has a return value
1182013702 M * Bertl same happens with case (on older gcc)
1182013718 M * daniel_hozac ah, that's weird...
1182013733 M * Bertl yeah, should be no problem with newer ones though
1182013785 M * Bertl http://paste.linux-vserver.org/2589
1182013840 Q * Wonka Quit: reboot. bbl.
1182013890 M * daniel_hozac i guess the {}s aren't really needed, but they do help some with the readability.
1182013937 M * Bertl btw, I think we need some space specific tests soon
1182013951 M * Bertl i.e. unsharing and entering with this or that space
1182014018 M * daniel_hozac indeed.
1182014102 M * daniel_hozac aren't we putting the same nsproxy twice there?
1182014114 M * Bertl where?
1182014122 M * daniel_hozac delta-nsproxy-fix04
1182014140 M * daniel_hozac won't both of them refer to p->nsproxy as it was before the unshare?
1182014215 M * daniel_hozac (i think it's right, i'm just confused by the differing names)
1182014240 M * Bertl yeah, we had the same with the mainstream function, sec
1182014322 M * Bertl what I'm wondering is, IMHO if we call sys_unshare() then an nsproxy should be leaked ...
1182014358 M * Bertl as we are basically doing exactly what sys_unshare() does (reagrding proxies)
1182014452 M * Bertl but I haven't testet that yet, have to fix up vcmd first :)
1182014489 M * daniel_hozac yeah, sys_unshare should leak the old nsproxy...
1182014528 M * Bertl that is my impression :)
1182015096 M * daniel_hozac what do you think the get_nsproxy is supposed to do in unshare_nsproxy_namespaces?
1182015262 M * Bertl you mean, why the hell do they take the reference?
1182015266 M * daniel_hozac yeah
1182015282 M * daniel_hozac i can't think of any valid reason.
1182015284 M * Bertl I think it is, because it happened that the 'old' nsproxy vanished at some point
1182015295 M * Bertl when being replaced without locking
1182015295 M * daniel_hozac to protect against races?
1182015299 M * Bertl yes
1182015316 M * daniel_hozac but shouldn't _that_ nsproxy then be passed on to the subsequent functions?
1182015318 M * Bertl it looks very hackish to me and I do not really like it
1182015320 M * daniel_hozac or returned to the caller?
1182015325 M * daniel_hozac or _something_ sane.
1182015341 M * Bertl yeah, probably, but let's see where it is used ...
1182015342 M * daniel_hozac because right now, if it is replaced, there's no way to get the old one...
1182015372 M * Bertl ah, no, the function itself is never used to _replace_ them
1182015430 M * Bertl actually it is only used once in mainline
1182015435 M * Bertl in sys_unsahre()
1182015444 M * Bertl and there it probably causes a leakage :)
1182015456 M * Bertl so much for testing before inclusion ...
1182015552 M * daniel_hozac well, i mean, if there is a race and another process changes the nsproxy.
1182015589 M * daniel_hozac making the old_ns in unshare_nsproxy_namespaces not point to the current nsproxy anymore.
1182015600 M * Bertl I think the race comes from the funny filesystem interface to change the spaces
1182015611 M * daniel_hozac ah... yeah, that would make sense.
1182015627 M * Bertl I don't see a reason to lock current in unsahre for example
1182015649 M * Bertl and the nsproxy exchange could as well be done with xchg :)
1182015655 M * daniel_hozac yeah.
1182018527 M * Bertl yep, confirmed, sys_unshare leaks a proxy ...
1182018611 M * Bertl it's not immediately obvious, even with accounting
1182018615 M * daniel_hozac hehe, nice.
1182018625 M * Bertl as it increments the ref to the 'main' proxy
1182018636 M * Bertl but with vcmd 0.09, you can easily do:
1182018642 M * Bertl vcmd -u ^17 -- vcmd -u ^17 -- sleep 10
1182018654 M * Bertl which will leak the second proxy too :)
1182018658 M * daniel_hozac hehe, okay.
1182018722 M * Bertl or even: vcmd -nu ^17 -- vcmd -nu ^17 -- sleep 10
1182018733 M * Bertl to rule out Linux-VServer stuff :)
1182018900 J * Wonka produziert@chaos.in-kiel.de
1182019198 M * Bertl wb Wonka!
1182019415 M * Wonka re
1182019422 M * Wonka vserver took some time to reboot...
1182019433 M * Wonka ext3 decided to have some issues
1182020145 J * _jthm_ ~infowolfe@c-24-10-234-129.hsd1.ut.comcast.net
1182020232 M * Bertl which kind of issues?
1182020546 Q * infowolfe_ Ping timeout: 480 seconds
1182020549 M * Wonka something with duplicate blocks
1182020677 J * infowolfe ~infowolfe@c-24-10-234-129.hsd1.ut.comcast.net
1182020833 M * Bertl daniel_hozac: can one wait on a sys_clone(2) process?
1182020939 Q * _jthm_ Ping timeout: 480 seconds
1182020942 M * Bertl ah, found it __WCLONE :)
1182021050 J * infowolfe_ ~infowolfe@c-24-10-234-129.hsd1.ut.comcast.net
1182021308 J * _jthm_ ~infowolfe@c-24-10-234-129.hsd1.ut.comcast.net
1182021373 N * DoberMann DoberMann[PullA]
1182021449 Q * infowolfe Ping timeout: 480 seconds
1182021686 Q * infowolfe_ Ping timeout: 480 seconds
1182021905 J * sladen paul@starsky.19inch.net
1182022265 M * HeinMueck41 is there some kind of document that describes what happens when you start a vserver with vserver start in a new config environment? I went through most of the stuff in the wiki, but still missing the big picture. Sorry, slow mind :)
1182022347 M * daniel_hozac vi /usr/lib*/util-vserver/vserver.start ;)
1182022380 M * HeinMueck41 yeah, that one I read, too :)
1182022439 M * daniel_hozac so, what are you missing?
1182022545 J * infowolfe_ ~infowolfe@c-24-10-234-129.hsd1.ut.comcast.net
1182022572 M * Bertl daniel_hozac: I'm still in the dark regarding the suse issue ... could you clarify that for me?
1182022600 M * daniel_hozac i'm in the dark too ;)
1182022608 M * Bertl I would assume that sysv init style sets the environment for the runlevel script correctly, no?
1182022614 M * HeinMueck41 A reason for a wiki? If you have code for documentation - why a wiki? ;-)
1182022625 M * Bertl i.e. including RUNLEVEL and PREVLEVEL, right?
1182022638 M * daniel_hozac Bertl: no, IIRC it leaves that to the runlevel scripts themselves.
1182022652 M * Bertl that is a bug then
1182022678 M * Bertl the documentation for init states clearly what should be set in the env
1182022689 M * HeinMueck41 right :)
1182022705 M * HeinMueck41 and the scripts are using it
1182022731 M * HeinMueck41 so if you do a shutdown or runlevel change, it tries to stop the services for PREVLEVEL
1182022748 M * HeinMueck41 if that is empty, it falls back to some default stuff
1182022758 M * HeinMueck41 which might, but must not meet what you need
1182022777 M * HeinMueck41 debian does that too
1182022789 M * HeinMueck41 but it does not depend on it that much as suse
1182022807 M * daniel_hozac PREVLEVEL is set on stop.
1182022862 M * HeinMueck41 both are always set before init runs any script 
1182022866 M * Bertl is there a problem with setting the init env according to the manpage?
1182022873 M * daniel_hozac no.
1182022921 M * Bertl okay, then doing so should instantly fix suse guests, no?
1182022948 M * HeinMueck41 without changing the rc script, yes
1182022973 M * daniel_hozac perhaps. the fact that plain doesn't work either though, suggests there are deeper issues to me...
1182022979 Q * _jthm_ Ping timeout: 480 seconds
1182022982 M * HeinMueck41 but is there a way to set the environment through scripts for the shell that calls the rc script?
1182023008 M * HeinMueck41 Thats why I asked about the process
1182023020 M * daniel_hozac you've already seen apps/init/environment, no?
1182023024 M * HeinMueck41 faking what init does before it runs rc
1182023026 M * HeinMueck41 yep
1182023037 M * HeinMueck41 but that is too static for the job
1182023053 M * Bertl daniel_hozac: yes, but he would have to change that in pre/post scripts every time
1182023053 M * HeinMueck41 it always set the same values 
1182023073 M * Bertl daniel_hozac: which IMHO is a real hack to work around a bug in util-vserver :)
1182023098 M * daniel_hozac HeinMueck41: it's also only used for startup :)
1182023102 M * daniel_hozac Bertl: sure.
1182023160 Q * onox Ping timeout: 480 seconds
1182023370 M * daniel_hozac http://people.linux-vserver.org/~dhozac/p/uv/experimental/delta-sysv-env-fix01.diff
1182023388 M * daniel_hozac does that do the trick?
1182023673 Q * sladen Ping timeout: 480 seconds
1182023817 M * HeinMueck41 yep, levels are set by that :)
1182023853 M * daniel_hozac and there are no more problems?
1182023867 M * HeinMueck41 suse likes to use the CONSOLE variable offered by init, too .
1182023871 M * HeinMueck41 oh, there are 
1182023879 M * HeinMueck41 of course :)
1182023896 M * HeinMueck41 I really wonder why it ever worked - hehe
1182023902 J * _jthm_ ~infowolfe@c-24-10-234-129.hsd1.ut.comcast.net
1182023936 M * HeinMueck41 it is funny that when you enter the vserver, you see only the processes you wanted to start
1182023998 M * HeinMueck41 but the hostname is the one of the root, the environment is the basic stuff from the host, just a few more things added
1182024027 J * infowolfe ~infowolfe@c-24-10-234-129.hsd1.ut.comcast.net
1182024298 J * sladen paul@starsky.19inch.net
1182024299 Q * infowolfe_ Ping timeout: 480 seconds
1182024389 Q * _jthm_ Ping timeout: 480 seconds
1182024411 M * HeinMueck41 startup and shutdown now works out of the box
1182024413 J * infowolfe_ ~infowolfe@c-24-10-234-129.hsd1.ut.comcast.net
1182024459 M * HeinMueck41 how is the environment injected into the vserver context? For instance, the hostname?
1182024487 M * daniel_hozac hmm?
1182024512 M * HeinMueck41 is the problem on my site, i.e. stupid configuration or does the guest system need to do anything to enable it?
1182024525 M * daniel_hozac to enable what, exactly?
1182024586 M * HeinMueck41 the hostname
1182024616 J * _jthm_ ~infowolfe@c-24-10-234-129.hsd1.ut.comcast.net
1182024634 M * HeinMueck41 if I could trace why the hostname of the guest is not the name configured, but the name of the root server, I might be able to find out what I'm doing wrong ...
1182024829 Q * infowolfe Ping timeout: 480 seconds
1182024833 M * daniel_hozac are you sure the guest doesn't set the hostname itself?
1182024995 M * HeinMueck41 so far - yes
1182025006 M * HeinMueck41 it would have a different name then
1182025019 Q * infowolfe_ Ping timeout: 480 seconds
1182025026 M * HeinMueck41 but it calls itself marvin, just like it's root
1182025054 M * HeinMueck41 $HOSTNAME is also set to marvin inside the guest
1182025068 Q * meetra 
1182025219 M * daniel_hozac and hostname also says marvin?
1182025229 M * daniel_hozac you configured it in /etc/vservers/<guest>/uts/nodename, right?
1182025296 M * HeinMueck41 hehe
1182025302 M * HeinMueck41 no, of course not :)
1182025312 M * HeinMueck41 just name
1182025336 M * HeinMueck41 done it - does it
1182025673 J * weasel_ weasel@asteria.debian.or.at
1182025704 N * weasel_ Guest2080
1182025783 Q * weasel Remote host closed the connection
1182025783 N * Guest2080 weasel
1182026187 M * HeinMueck41 works just fine now
1182026212 M * HeinMueck41 Nevertheless, I had to remove two lines from the rc 
1182026214 M * HeinMueck41 if test "$PREVLEVEL" = "N" ; then
1182026214 M * HeinMueck41 #    test -e /proc/cmdline || mount -n -t proc proc /proc
1182026214 M * HeinMueck41 #    read -t 2 check < /proc/cmdline 2> /dev/null
1182026235 J * infowolfe_ ~infowolfe@c-24-10-234-129.hsd1.ut.comcast.net
1182026237 M * daniel_hozac so you want PREVLEVEL=-?
1182026257 M * HeinMueck41 I think N is fine
1182026284 M * HeinMueck41 but what suse does with it is okay, but troublesome in a guest
1182026301 M * HeinMueck41 one just need to know it
1182026519 M * daniel_hozac having to edit the rc scripts is bad though.
1182026602 M * HeinMueck41 the only way to get around it is to put something into /proc/cmdline that disables the goal of these two lines
1182026639 Q * _jthm_ Ping timeout: 480 seconds
1182026640 M * daniel_hozac unfortunately, /proc/cmdline is hidden for a reason.
1182026654 M * HeinMueck41 thought so, yes
1182026695 M * daniel_hozac and while there are interests in virtualizing that, that's not in any vanilla patch right now...
1182026865 J * _jthm_ ~infowolfe@c-24-10-234-129.hsd1.ut.comcast.net
1182027289 Q * infowolfe_ Ping timeout: 480 seconds
1182027297 J * infowolfe ~infowolfe@c-24-10-234-129.hsd1.ut.comcast.net
1182027656 Q * _jthm_ Ping timeout: 480 seconds
1182028160 J * meetra ~xtreme@82.155.150.227
1182029926 J * meandtheshell ~markus@85-124-232-67.work.xdsl-line.inode.at
1182031102 M * Bertl I think we can attack virtualizing cmdline soon ...
1182031559 J * Aiken ~james@ppp121-45-250-169.lns2.bne4.internode.on.net
1182032272 M * HeinMueck41 Just one last question 
1182032289 M * HeinMueck41 In the suse rc there is one line saying
1182032292 M * HeinMueck41 test -z "$CONSOLE" && CONSOLE=/dev/console
1182032314 M * HeinMueck41 which leads to output directed to /dev/console
1182032331 M * HeinMueck41 the flower pages mentions a tty link in apps/init
1182032358 M * daniel_hozac CONSOLE=/dev/tty in apps/init/environment?
1182032358 M * HeinMueck41 would that overwrite the rc stuff?
1182032383 M * HeinMueck41 because $CONSOLE is not set I do not get any output from the start/stop scripts
1182032496 M * HeinMueck41 ah, sure
1182032503 M * HeinMueck41 let me see
1182032983 M * HeinMueck41 hm, that is removed when rc finishes, so it just like the runlevel stuff gone when you stop the system
1182033242 Q * Medivh Ping timeout: 480 seconds
1182033271 M * HeinMueck41 I just added it at the same place where you put the PREVLEVEL stuff and that works
1182033921 J * Medivh ck@paradise.by.the.dashboardlight.de
1182034265 Q * dna Quit: Verlassend
1182034559 Q * HeinMueck41 Quit: Aah!
1182034693 J * DavidS david@chello062178045213.16.11.tuwien.teleweb.at
1182035046 M * DavidS hi Bertl, folks!
1182035388 M * daniel_hozac hello DavidS