1188864054 M * igraltist !status
1188864147 M * Bertl wrong channel? :)
1188864171 M * igraltist oh :)
1188864657 N * DoberMann DoberMann[ZZZzzz]
1188864811 M * Bertl daniel_hozac: btw, the best idea I came up with (regarding the initpid) is to use a tagging on the dentry to detect entries from the 'wrong' context
1188864826 M * Bertl (in case you want to try something in this direction)
1188864943 J * coderanger_ ~coderange@c-65-96-210-168.hsd1.ma.comcast.net
1188864992 M * daniel_hozac how would the dentry tag be different from the inode tag/task tag?
1188865036 J * jonathan ~chatzilla@host-181-96.dsl-sea.seanet.com
1188865084 M * Bertl hmm, good point, it should (theoretically) suffice to check for the task
1188865379 M * jonathan We've successfully built the vserver kernel for FC6 but are stumped when it comes to building a package to be used for building guests.  How do *WE* build our own guest package based on your system?
1188865405 M * jonathan Oops.  Make that 'based on *our* system'.
1188865416 M * daniel_hozac hmm?
1188865423 M * Bertl hmm, I think util-vserver should be able to do so, no?
1188865614 M * daniel_hozac btw, you know FC6 is just a few months from EOL, right? 
1188865688 M * jonathan We're aware of that but need to start somewhere.  So far I've seen documentation that talks about using dump | restore to create a guest.  Does vserver provide any tools to help with this?
1188865718 M * daniel_hozac vserver - build --help
1188865721 M * Bertl again, util-vserver :)
1188865776 M * daniel_hozac typically you'd just use yum/debootstrap/whatever's appropriate for your distro to net-install a guest...
1188865827 M * Supaplex any new ideas for http://rafb.net/p/Xgg16B57.html ?
1188865998 M * daniel_hozac why doesn't your guest have /var/run/utmp? :)
1188866021 M * Supaplex it's openwrt. I crated an empty one, and had a similar issue with chroot-sh
1188866027 M * Supaplex http://rafb.net/p/At2C8j46.html
1188866063 M * daniel_hozac Bertl: i seem to be able to separate the two inits now...
1188866068 M * Supaplex result is also the same if I copy the host utmp there.
1188866085 M * daniel_hozac Supaplex: do you have some mounts going on?
1188866099 M * Supaplex nope. it was created with -m skeleton
1188866231 M * Bertl daniel_hozac: excellent, that should do the trick then
1188866256 M * daniel_hozac it's... i don't think ugly suffices.
1188866283 M * jonathan So could I use 'vserver ~ build -m rsync ...' to copy over the parts of the filesystem I want as long as I avoid (/dev, /etc, /proc)?
1188866294 M * Supaplex how does -m rsync differ from -m skeleton and doing rsync myself?
1188866297 M * Bertl daniel_hozac: cruel abomination of a hack? :)
1188866342 M * daniel_hozac it doesn't :)
1188866347 M * daniel_hozac Bertl: something like that...
1188866516 M * Bertl you need to do two things, a) reinstate the old revalidate code
1188866532 M * Bertl and b) drop the dentry when it is being looked up
1188866562 M * daniel_hozac b doesn't seem to be working.
1188866634 M * daniel_hozac even though i added an if (vx_current_initpid(tgid)) goto out; in proc_pid_lookup.
1188866679 M * daniel_hozac http://people.linux-vserver.org/~dhozac/p/k/delta-initpid-fix03.diff
1188866717 M * daniel_hozac i suppose the revalidation could be simplied to just name_to_int != vx_map_pid
1188866816 M * Bertl nah, proc_pid_lookup is never called, IIRC
1188866832 M * Bertl it happens in the dentry lookup (dentry cache)
1188866927 M * daniel_hozac could we fix that part by making sure vx_find_proc_task_by_pid doesn't return init for the real pid?
1188866941 M * daniel_hozac or, nah.
1188866951 M * Bertl no, been there, done that, didn't work
1188866990 M * daniel_hozac hmm, shouldn't name_to_int != vx_map_pid in revalidate work for that too?
1188867050 M * Bertl the problem is the following case: guest with init pid virtualization looks up init 1->dentry(1)->init of guest (!=1)
1188867066 M * Bertl then the host looks up init 1->dentry_cache(1)
1188867119 M * daniel_hozac isn't this the case i've fixed?
1188867142 M * Bertl you tell me?
1188867155 M * daniel_hozac vps shows both inits separately now.
1188867171 M * Bertl the problem is, that there will be no pid_lookup or revalidate
1188867207 M * Bertl separately?
1188867218 M * daniel_hozac i.e. the host's init and the guest's init.
1188867219 M * Bertl you mean, you get '1' twice in readdir?
1188867223 M * daniel_hozac no.
1188867228 M * Bertl so?
1188867240 M * daniel_hozac before, without the revalidate hunk, i got the guest's init twice.
1188867305 M * Bertl please, if your test, always test both, readdir and the actual path walk
1188867328 M * Bertl because they are completely independant (unfortunately)
1188867332 M * daniel_hozac so readdir is the problem here, right?
1188867340 M * Bertl and the readdir part is trivial to fix, but the pathwalk not
1188867368 M * Bertl for the readdir we do not even need to validate the proper init :)
1188867389 M * Bertl it is sufficient to show _an_ init (whatever init it is as pid=1)
1188867398 M * daniel_hozac right...
1188867424 M * daniel_hozac well, i'm pretty sure i've got the pathwalk down. i'm rebooting now to test the /proc/<real pid> fix.
1188867433 M * Bertl the thing you basically want to test for is the following:
1188867450 M * Bertl ls /proc/1/status on host
1188867456 M * Bertl s/ls/cat/
1188867465 M * Bertl then the same on the virtualized guest 
1188867472 M * Bertl and again back on the host
1188867495 M * daniel_hozac right.
1188867506 M * daniel_hozac i'm pretty sure that already works.
1188867525 M * Bertl okay, so what doesn't work then?
1188867767 M * daniel_hozac hang on, i broke it :)
1188867885 P * jonathan 
1188868235 M * daniel_hozac http://people.linux-vserver.org/~dhozac/p/k/delta-initpid-fix04.diff seems to be working fine, and i don't see any issues with it.
1188868340 M * daniel_hozac i.e. /proc/1 gets me the right pid every time, /proc/<real pid> is no longer accessible in the guest.
1188868542 M * Bertl okay, please double check _all_ cases
1188868582 M * Bertl i.e. host and 3 guest setups, and move between all of them for readdir and blind /proc/1 /proc/<pid> tests
1188868900 M * daniel_hozac everything seems to be behaving as expected.
1188868912 M * Bertl excellent work then!
1188868977 M * Supaplex good patch. have a biscuit.
1188869004 M * Bertl excellent one, have a dozen ... :)
1188869013 M * daniel_hozac haha.
1188869057 M * daniel_hozac the vx_find_proc_task_by_pid does make me somewhat nervous though. it seems like we would've used real_pid for a reason there.
1188869105 M * daniel_hozac but since i can't think of any reason right now, i'm not going to worry about it :)
1188869472 M * daniel_hozac umm, yeah. there's a put_task_struct missing there.
1188869512 Q * coderanger_ Quit: coderanger_
1188869761 M * daniel_hozac http://people.linux-vserver.org/~dhozac/p/k/delta-initpid-fix05.diff is the version i'm sticking with.
1188869761 Q * FireEgl Read error: Connection reset by peer
1188870119 J * DoberMann_ ~james@AToulouse-156-1-57-79.w90-16.abo.wanadoo.fr
1188870227 Q * DoberMann[ZZZzzz] Ping timeout: 480 seconds
1188870286 M * Supaplex sigh...
1188870551 M * Supaplex woah. utmp was deleted (before, and now), but now it's yelling something completly different. http://rafb.net/p/F8vWlm77.html
1188870611 M * daniel_hozac your guest doesn't support the sysv initstyle that you've chosen.
1188870838 M * daniel_hozac Bertl: okay, something is definitely wrong with it... will investigate.
1188870966 M * Bertl ok
1188871017 M * Supaplex is sh -l a bashism?
1188871040 M * daniel_hozac stopping the guests causes a BUG (seems to be bad ref counting), eventually it leads to a totally unusable system...
1188871377 Q * ensc Remote host closed the connection
1188872339 M * Supaplex http://riddle.daxal.com/~supaplex/vserver+openwrt/htop.png
1188872343 M * Supaplex yay
1188872371 M * Bertl solved?
1188872381 M * Supaplex yup
1188872391 M * Supaplex I don't really know how though :P
1188872505 Q * ntrs__ Ping timeout: 480 seconds
1188872538 M * Bertl do you use xid/tagging?
1188872553 M * Supaplex not afaik
1188872566 M * Bertl so not mounted with tag or tagxid
1188872600 M * Supaplex just defaults
1188872635 M * Bertl then it probably just was an userspace issue
1188872666 M * Supaplex I don't doubt it. maybe I'll try to replicate the issue on a new guest later.
1188872679 M * Bertl yeah, would definitely help
1188872727 M * Bertl I'm not a fan of mysteriously resolved issues :)
1188872859 M * Supaplex me either
1188872942 J * virtuoso_ ~s0t0na@pppoe-5.58.110.89-adsl.spbnit.ru
1188873352 Q * virtuoso Ping timeout: 480 seconds
1188874741 Q * quasisane Quit: ERC Version 5.2 (IRC client for Emacs)
1188876167 M * Bertl okay, off to bed .. have a good one, everyone!
1188876172 N * Bertl Bertl_zZ
1188876399 J * FireEgl FireEgl@4.0.0.0.1.0.0.0.c.d.4.8.0.c.5.0.1.0.0.2.ip6.arpa
1188876463 A * Supaplex tries as he may to rid the jungle of some excess cables
1188878971 J * coderanger_ ~coderange@c-65-96-210-168.hsd1.ma.comcast.net
1188881309 J * igraltista ~jens@pD9E4FC46.dip.t-dialin.net
1188881320 M * igraltista hi
1188881368 M * igraltista when i play a mp3 from a vguest wich offers it over samba then the song is always brocken
1188881389 M * igraltista if i download it local and play it all is fine
1188881420 M * igraltista so something in the vserver has to less priority mayby
1188881786 J * balbir ~balbir@194.42.125.16
1188881815 M * daniel_hozac igraltista: i watch movies over samba shares on guests on a regular basis...
1188881868 M * igraltista is possibly that encryption make it broken
1188882058 M * igraltista hmm i can copy in the same time a movie and noting are disturbed, but then again it break always
1188882076 M * daniel_hozac "break"?
1188882094 M * igraltista short break in the song 
1188882107 M * igraltista i dont know a english word for this
1188882150 M * daniel_hozac so movies work fine?
1188882280 M * igraltista no i did copy while listen to mp3 song a movie and this not disturbed 
1188882282 Q * balbir Ping timeout: 480 seconds
1188882304 M * igraltista so the harddrive should haveenough power for work i hope
1188883473 M * igraltista so i try movie to watch and it not break form time to time
1188883714 Q * igraltista Read error: Connection reset by peer
1188883757 J * balbir ~balbir@194.42.125.16
1188884494 J * Pazzo ~ugelt@sadsl-246059.rol.raiffeisen.net
1188884983 Q * Pazzo Quit: ...
1188885428 N * DoberMann_ DoberMann
1188885579 J * sharkjaw ~gab@158.36.44.106
1188886087 Q * mnemoc Ping timeout: 480 seconds
1188886570 J * mnemoc ~amery@kilo105.server4you.de
1188887283 J * quasisane ~sanep@c-76-118-191-64.hsd1.nh.comcast.net
1188887650 N * DoberMann DoberMann[PullA]
1188887675 J * yarihm ~yarihm@84-74-20-183.dclient.hispeed.ch
1188887784 J * ensc ~irc-ensc@p54B4FBE4.dip.t-dialin.net
1188889688 J * dna ~dna@29-234-dsl.kielnet.net
1188890102 J * sasaki_ta ~sasaki_ta@nttkyo767180.tkyo.nt.ftth.ppp.infoweb.ne.jp
1188890127 M * sasaki_ta hello
1188890162 M * sasaki_ta I am try vserver on debian etch.
1188890233 M * sasaki_ta I want to copy a guest, and run it another root-server.
1188890297 M * sasaki_ta vserver-copy is not work. and I am trying vserver build -m rsync.
1188890328 M * arachnist vserver foobar build clone ?
1188890345 M * sasaki_ta Is it working? I got error "# vserver test001 -m rsync --interface test001=eth0:19.168.1.190 --hostname test001.local.colt.co.jp  -- --source hawai
1188890346 M * sasaki_ta Can not find a vserver-setup at '/etc/vservers/test001/'."
1188890376 M * daniel_hozac vserver test0001 build ....
1188890388 M * sasaki_ta woooo
1188890394 M * sasaki_ta thank you!
1188890986 M * sasaki_ta OK. done.
1188891044 M * sasaki_ta I will post blog. for japanese.
1188891261 M * arachnist how do i make a vserver start with linux32, without typing it everytime?
1188891287 M * daniel_hozac echo linux32 > /etc/vservers/<guest>/personality
1188891293 M * arachnist thanks
1188891480 Q * yarihm Quit: This computer has gone to sleep
1188893485 A * Supaplex ponders sleep()
1188893779 J * friendly12345 ~friendly@ppp59-167-82-199.lns2.mel6.internode.on.net
1188894264 Q * rorem-78 Remote host closed the connection
1188894504 J * rorem- ~roremtank@bzq-219-46-202.isdn.bezeqint.net
1188898262 J * MooingLemur ~troy@shells195.pinchaser.com
1188900691 J * Piet ~piet@tor.noreply.org
1188900811 J * meandtheshell ~markus@85.127.114.89
1188902380 J * yarihm ~yarihm@whitehead2.nine.ch
1188902400 Q * yarihm 
1188902467 J * yarihm ~yarihm@whitehead2.nine.ch
1188903155 J * onox ~onox@kalfjeslab.demon.nl
1188903206 J * Julius ~julius@p57B277A8.dip.t-dialin.net
1188903960 Q * Julius Ping timeout: 480 seconds
1188904215 Q * Piet Ping timeout: 480 seconds
1188904251 J * Piet ~piet@tor.noreply.org
1188904623 J * Julius ~julius@p57B277A8.dip.t-dialin.net
1188904656 Q * sasaki_ta Quit: Computer goes to sleep!
1188905168 J * emtt1 ~eric@dynamic-acs-24-154-85-144.zoominternet.net
1188907482 Q * Piet Remote host closed the connection
1188907552 J * Piet ~piet@tor.noreply.org
1188908288 Q * Julius Ping timeout: 480 seconds
1188908433 J * Julius ~julius@p57B277A8.dip.t-dialin.net
1188908846 J * fliegerhermi ~5b1041c2@pc-intern.com
1188908893 M * fliegerhermi Hi guys... I got a question: What's the best way to copy a vserver-guest to another host? I guess rsync? Do you copy the host up and running or do you shut it down before copying?
1188908923 M * daniel_hozac first while it's running, then stop it, rsync the final bits, start it on the other host.
1188908932 M * [Guy] fliegerhermi: I'd suggest taking an LVM snapshot and rsyncing that, if avoiding downtime is important
1188908945 N * [Guy] Guy-
1188908952 M * fliegerhermi no downtime is not a problem
1188908957 M * fliegerhermi got backupservers...
1188908977 M * fliegerhermi so I rsync twice @daniel? Same command? :-?
1188908977 Q * balbir Ping timeout: 480 seconds
1188908983 M * daniel_hozac yes.
1188908990 M * fliegerhermi alright! Thanks!
1188909001 M * daniel_hozac of course, if you don't care about downtime, it doesn't matter.
1188909009 M * daniel_hozac just stop the guest, rsync it once, start it on the other host.
1188909053 M * fliegerhermi well that sounds easier.... 
1188909095 M * Hollow and don't forget to use --numeric-ids :)
1188909103 M * fliegerhermi I just remembered that I had problems the last time. So I thaught I just ask..... Last time it didn't work with a running / stoped guest. So I just wanted to doublecheck. Thanks!
1188909110 M * fliegerhermi Thanks ;-)
1188909397 P * friendly12345 
1188909742 M * fliegerhermi rsync --verbose --compress --recursive --stats --links --numeric-ids -e ssh /vserver/guest root@myhost_new:/vserver/guest
1188909749 M * fliegerhermi that should work fine right?
1188909789 M * daniel_hozac if you _really_ use /vserver, sure.
1188909839 M * fliegerhermi vservers
1188909847 M * fliegerhermi right
1188909851 M * fliegerhermi ok... Thanks a lot
1188910230 J * balbir ~balbir@194.42.125.16
1188910330 Q * igraltist Ping timeout: 480 seconds
1188910777 Q * fliegerhermi Quit: CGI:IRC (Ping timeout)
1188911327 J * marcfiu ~mef@aegis.CS.Princeton.EDU
1188911336 P * marcfiu 
1188911710 Q * balbir Ping timeout: 480 seconds
1188911838 Q * sharkjaw Quit: Leaving
1188912901 J * pmenier ~pmenier@LNeuilly-152-22-72-5.w193-251.abo.wanadoo.fr
1188913238 M * Hollow daniel_hozac: ping
1188913512 J * hipe ~hipe@BSN-77-69-204.dsl.siol.net
1188913608 Q * hipe 
1188914050 N * Bertl_zZ Bertl
1188914057 M * Bertl morning folks!
1188914089 Q * Aiken Remote host closed the connection
1188914181 M * Hollow morning Bertl 
1188914205 M * Bertl how's going?
1188914232 M * Hollow good, still enjoying vacation :)
1188914242 M * Hollow well, still a month to go :)
1188914651 J * Aiken ~james@ppp121-45-255-55.lns2.bne4.internode.on.net
1188914754 M * Bertl wb Aiken!
1188914826 M * Aiken hi
1188914871 Q * Julius Remote host closed the connection
1188915278 Q * Aiken Quit: Leaving
1188915449 Q * quasisane Ping timeout: 480 seconds
1188915700 Q * zbyniu Ping timeout: 480 seconds
1188916129 J * Val ~val@v41.org
1188916139 M * Val Hi
1188916162 J * hallyn ~xa@adsl-75-0-147-61.dsl.chcgil.sbcglobal.net
1188916215 M * Val did someone test nstx (Bogus DNS server, for IPv4 over DNS tunnels) under a guest ?
1188916264 M * Val i got a problem to access tun/tap under ntsxd dedicated vserver guest
1188916340 M * Val i made tunO/tap0 persistant with tunctl on the host, made available the both under the guest like explained for OpenVPN in the FAQ 
1188916457 M * Bertl okay, sounds good so far?
1188916469 M * Bertl did you assign an ip to the tun device?
1188916518 M * Val no, the IP is assigned by util-vserver when starting the guest
1188916544 M * Bertl okay, sounds good, does the nstx connect to the proper tun device?
1188916552 M * Bertl (maybe it tries to open a new one?)
1188916616 M * Val no, i got "Tun (/dev/net/tun): /dev/net/tun not found." and "/dev/tap0 not found." 
1188916627 M * Val and it faild
1188916630 M * Val fails
1188916646 M * Bertl well, /dev/net/tun is the multiplexor
1188916654 M * Bertl you might want to copy that into the guest
1188916678 M * Val it's already done
1188916698 M * Bertl then it seems strange that it says: /dev/net/tun not found.
1188916706 M * Val ls -laF /dev/net/tun /dev/tap0
1188916706 M * Val crw-rw---- 1 root adm 10, 200 2007-09-04 15:48 /dev/net/tun
1188916706 M * Val crw------- 1 root adm 36,  16 2007-09-04 16:06 /dev/tap0
1188916712 M * Val yup
1188916717 M * Bertl maybe use strace -fF on it?
1188916720 Q * Piet Ping timeout: 480 seconds
1188916725 M * Val yes, i'm installing strace
1188916741 M * Bertl and remove the tap0, it is just wrong :)
1188916743 M * sid3windr crw-r--r-- 1 root root 10, 200 Mar  4  2005 /dev/net/tun
1188916753 M * sid3windr my vpn box
1188916768 M * Bertl yeah, if it runs as user, o+r could help too
1188916779 M * Val i test it under root 
1188916787 M * sid3windr your configfile doesn't have a user statement?
1188916791 M * Val no
1188916793 M * sid3windr k
1188916799 M * sid3windr strace then
1188916806 M * Val i try it manualy first
1188916811 M * Val with debug set :)
1188916842 J * Piet ~piet@tor.noreply.org
1188916898 M * Val Bertl: why should i remove the tap0 file ? 
1188916918 M * Val what is wrong ?
1188916935 M * Bertl well, first, tap devices are for ethernet level traffic
1188916948 M * Bertl I presume you want to use IP level (tun) devices
1188916966 M * Val yup
1188916973 A * sid3windr tapdances
1188916975 M * Bertl second they are 'network' devices, so they do not live in /dev
1188916996 A * sid3windr wonders why openvpn looked for /dev/tap0 then ;)
1188917083 M * Val http://zbla.net/strace-nstxd.txt
1188917170 M * Bertl sid3windr: why does bind9 need a capability to drop it?
1188917179 M * sid3windr drop what? :)
1188917269 M * Bertl Val: well, another example of useless debug output generated by clueless developers (it seems)
1188917283 M * Val :)))
1188917302 M * Bertl inside the guest, do you see the tun device?
1188917324 M * Val yup, and it is configured 
1188917336 M * Bertl does it have a MAC address?
1188917362 M * Val yes
1188917372 M * Bertl then it isn't a tun device :)
1188917378 M * Val ok :)
1188917393 M * Bertl so, you want to remove that tap device (which is labeled tun :)
1188917400 M * sid3windr should be 00-00-00-00-00-00-00-00-00-00-00-00-00-00-00-00 ;)
1188917401 M * Bertl and create a tun device instead :)
1188917428 M * Val ok
1188917577 M * zLinux hello
1188917651 M * Bertl hey zLinux!
1188917683 M * michal` zLinux sounds like a linux on zseries
1188917692 M * michal` what's really enjoyable :)
1188917724 M * harry Q: how can i loop mount an iso image into a running vserver
1188917749 M * michal` by entering into that vserver namespace and mounting it there
1188917806 M * harry spank you
1188918027 M * daniel_hozac Hollow: pong
1188918045 M * Val well... /sbin/ifconfig tun0 up 10.0.0.1 netmask 255.255.255.0 on the host brings up a tun0 interface with HW Max address :(
1188918095 M * Hollow daniel_hozac: for every nfs share on the host i get a warning umount /foo/bar: not mounted.. i did not have time to investigate yet
1188918096 M * sid3windr which is not 00-00..?
1188918105 M * Hollow daniel_hozac: on vserver ... start that is
1188918106 M * Val ... or an tun0: ERROR while getting interface flags
1188918113 M * Bertl Val: get a tool to create a tun device (you can find one on my storage area)
1188918121 M * Val ok
1188918151 M * daniel_hozac Hollow: that's... interesting.
1188918215 M * daniel_hozac Hollow: is there a top-level mount above them all?
1188918253 M * Val Bertl: http://vserver.13thfloor.at/Experimental/ ?
1188918264 M * Hollow well, i have two nfs mounts, both mounted to a directory belonging to /
1188918310 J * ktwilight ~ktwilight@102.123-66-87.adsl-dyn.isp.belgacom.be
1188918353 M * Bertl welcome ktwilight!
1188918373 M * Bertl http://vserver.13thfloor.at/Stuff/tunctl.tar.bz2
1188918398 M * Val ok thx
1188918490 J * zLinux_ ~zLinux@88.213.29.46
1188918514 Q * zLinux Ping timeout: 480 seconds
1188918580 M * daniel_hozac Hollow: hmm. that's very strange then.
1188918647 M * daniel_hozac Hollow: only NFS mounts too?
1188918652 M * Bertl Hollow: mtab?
1188918672 M * Hollow yes, only the nfs mounts, i really have lots of mounts on this machine, but only the 2 nfs give a warning
1188918678 Q * ktwilight_ Ping timeout: 480 seconds
1188918722 M * Hollow Bertl: seems to be correct in mtab
1188918754 J * cryptronic ~oli@p54A3B9D5.dip0.t-ipconnect.de
1188918756 M * cryptronic hi all
1188918757 M * Val thx Bertl, it works now, tunctl packaged for Debian create a level 2 tun/tap interface only
1188918769 M * Bertl Val: yes, I know, it is broken
1188918818 M * daniel_hozac Hollow: the weirdest part is that we read /proc/mounts to get the list of things to umount, so i don't see why it would fail...
1188918818 M * Bertl (IIRC, we discovered that back the time when I uploaded this tar)
1188918820 M * cryptronic do i need a special kernel header file when compiling vserver utils 214? I get the following error: http://paste.linux-vserver.org/6496
1188918831 M * Bertl cryptronic: debian?
1188918846 M * cryptronic yes
1188918890 M * Hollow daniel_hozac: yeah, strange indeed... i suppose it is related to the umount magic for the newly created namespace, right?
1188918902 M * Bertl daniel_hozac: seems they have finally updated all their kernel headers and such ...
1188918921 M * daniel_hozac Hollow: i assume it comes from the namespace cleanup, yeah.
1188918939 M * Hollow ok, i will debug it later
1188918960 M * Val Bertl: tun0 is now good, but i now get "Tun (/dev/net/tun): Permission denied. You usually have to be root to use nstx." But i'm running it as root :-/
1188919052 M * Bertl strace is your friend :)
1188919156 M * cryptronic is there a problem comiling uv 214 on debian?
1188919184 M * Bertl cryptronic: it seems so, but I would have expected you to hit the ext2 headers
1188919208 M * Bertl cryptronic: I'm sure daniel_hozac has some ideas here ...
1188919215 M * daniel_hozac it's not a problem compiling util-vserver 0.30.214.
1188919230 M * daniel_hozac _nothing_ has been changed in this area for a very long time, it's old compatibility code.
1188919244 M * daniel_hozac it's your e2fslibs headers breaking it.
1188919294 M * Bertl daniel_hozac: but the FS_IOC_GETFLAGS is new, no?
1188919296 M * cryptronic great seems to work quite fine now, after installing the header giles
1188919326 M * cryptronic /
1188919339 M * cryptronic /giles/files/ after installing the e2fslibs-dev
1188919377 M * daniel_hozac Bertl: right, and ext2_fs.h doesn't include fs.h.
1188919653 M * cryptronic is there a distribution script out there for becomming a 10.* suse guest?
1188919682 M * daniel_hozac not yet.
1188919893 M * Val c
1188919954 M * tokkee D'oh...
1188919964 A * tokkee has a hiligh on "s.h." ;-)
1188919977 M * tokkee ... like in "fs.h."... ;-9
1188920061 M * sid3windr :p
1188920072 M * sid3windr why?
1188920077 M * daniel_hozac Bertl: http://paste.linux-vserver.org/6497
1188920099 M * daniel_hozac Bertl: that's the bug i got with the initpid fix. i don't really see the relation though.
1188920170 M * Bertl okay, I should have my stuff together tonight, I'll look into that after finally getting out 2.3.0.18
1188920171 M * daniel_hozac dm-5 is my /vservers filesystem.
1188920180 M * daniel_hozac okay.
1188920230 Q * Piet Ping timeout: 480 seconds
1188920273 M * daniel_hozac i guess it's possible it's not due to the initpid fix at all...
1188920285 M * daniel_hozac i'll redo the same tests without it.
1188920300 M * Bertl okay, good idea, it happens on umount, yes?
1188920321 M * daniel_hozac i'm not sure what causes it, it happens when stopping the guest.
1188920338 M * daniel_hozac so probably implicit umount via namespace destruction.
1188920352 M * Bertl yeah, probably
1188920589 J * Piet ~piet@tor.noreply.org
1188920665 N * ensc Guest2666
1188920675 J * ensc ~irc-ensc@p54B4DE94.dip.t-dialin.net
1188920755 M * Val (bye & thx)
1188920758 P * Val 
1188920782 Q * Guest2666 Ping timeout: 480 seconds
1188920824 Q * FireEgl Read error: Connection reset by peer
1188921539 Q * yarihm Quit: Leaving
1188921592 M * cryptronic short question to centos5 guests ho do i set the $releasever correctly inside the guest so that this guest has the possibility of using yum
1188921631 J * zbyniu ~zbyniu@host13-188.crowley.pl
1188921633 M * daniel_hozac http://linux-vserver.org/Installation_on_CentOS
1188921643 M * daniel_hozac see the bottom.
1188921694 M * cryptronic ahh thanks this was missing there: http://linux-vserver.org/util-vserver:Setup_CentOS_Guest
1188921744 M * cryptronic i just added it at the bottom of that page
1188921776 J * FireEgl FireEgl@4.0.0.0.1.0.0.0.c.d.4.8.0.c.5.0.1.0.0.2.ip6.arpa
1188921823 M * daniel_hozac that description doesn't match what it does though...
1188922133 J * bonbons ~bonbons@2001:960:7ab:0:20b:5dff:fec7:6b33
1188922788 J * callahan ~chatzilla@host-181-96.dsl-sea.seanet.com
1188922905 P * callahan 
1188923178 J * jmcaricand ~jmcarican@d83-179-177-236.cust.tele2.fr
1188923351 M * Bertl wb bonbons, jmcaricand!
1188923375 M * jmcaricand wb Bert
1188923432 M * Bertl hmm, well, I don't remember leaving ... :)
1188923451 M * sid3windr :P
1188924507 M * micah daniel_hozac: forgot to tell you new util-vserver uploaded yesterday, will backport once it enters testing
1188924530 M * daniel_hozac yeah, i saw.
1188924533 M * Bertl micah: ah, with some workarounds?
1188924587 M * micah Bertl: yeah, i couldn't figure out the typedefs, but undef'ing __strict_ansi__ in the right spots was easier anyways :)
1188924618 M * Borg- heh.. interesting programming aproach..
1188924626 M * Bertl micah: and everything works fine now?
1188924659 M * micah Bertl: it compiled and I was able to do typical util-vserver tasks without fail, so AFAICT, yes
1188924698 M * Bertl good, I expect testme.sh and testfs.sh to be part of those typical tasks ...
1188924718 M * micah yes, although the wider debian audience will result in more broad test cases
1188924757 M * Borg- Bertl: howdy,, sup w/ that UDP loopback (127.0.0.1) ?
1188924762 M * Bertl micah: of course, note that newer testfs.sh has a slightly different option syntax (i.e. -xyz is advised)
1188924776 M * Borg- I was unable to setup newer VS.. ENOBOX.. I consumed them for other urgent stuff
1188924782 M * daniel_hozac note that this will require a 2.4 kernel to test...
1188924797 M * daniel_hozac at least.
1188924809 Q * pmenier Quit: pmenier
1188924839 M * Bertl Borg-: should work fine with 2.3.0.17+, please let us know if not
1188924874 M * Borg- okey :)
1188925131 M * cryptronic micah are you responsible for the debian package?
1188925150 J * Guest2678 ~ensc@www.sigma-chemnitz.de
1188925169 M * ard i
1188925185 M * ard hmpf...
1188925191 M * ard wrong window got focus
1188925210 M * Bertl np :)
1188925315 Q * coderanger_ Quit: coderanger_
1188925400 Q * Guest2678 Remote host closed the connection
1188926626 N * DoberMann[PullA] DoberMann
1188927314 J * quasisane ~user@c-76-118-191-64.hsd1.nh.comcast.net
1188928410 M * micah Bertl: I just noticed there was a newer testfs.sh, so I am going to run it now. 
1188928418 M * micah cryptronic: yes
1188928455 M * cryptronic micah maybe you could add a build dependency on e2fslibs-dev, because this lets my amd64 build breake
1188928512 M * micah cryptronic: in what way does it break?
1188928537 M * cryptronic it won't compile without the header files: http://paste.linux-vserver.org/6496
1188928657 M * micah cryptronic: ok, its added, will be in next upload
1188928666 M * cryptronic thanks a lot
1188928764 M * micah cryptronic: no problem!
1188928797 M * micah Bertl: ok, so now I should do testfs.sh -xyz -l -t -D /dev/loop0 -M /mnt 
1188928814 M * micah (this page should be updated? http://linux-vserver.org/Report_a_Bug to reflect this)
1188928839 Q * duckx Remote host closed the connection
1188928864 J * duckx ~Duck@tox.dyndns.org
1188928882 M * Bertl micah: for example, a few errors are expected, so better upload it for checking
1188928908 M * Bertl btw, -l is only for legacy IIRC, so newer kernels should not need it
1188928962 M * micah Bertl: http://paste.linux-vserver.org/6500
1188928971 M * micah yeah, I did both -l and without
1188929019 A * micah installs reiserfsprogs, xfsprogs and jfsprogs
1188929028 M * Bertl as you can see, ext2 and ext3 seems okay without legacy
1188929054 M * micah yeah
1188929068 M * daniel_hozac but the hack would only affect legacy.
1188929160 M * micah btw: http://svn.debian.org/wsvn/pkg-vserver/util-vserver/trunk/debian/patches/01strict_ansi.dpatch?op=file&rev=0&sc=0
1188929223 M * Bertl daniel_hozac: yeah, I wondered, are those headers included without legacy at all?
1188929254 M * daniel_hozac no.
1188929268 M * daniel_hozac they're all part of the fscompat API.
1188929299 A * micah re-runs with reiser/xfs/jfs programs installed
1188929360 M * Bertl daniel_hozac: i.c. and for what kernels do we need that?
1188929370 M * micah Bertl: updated tests: http://paste.linux-vserver.org/6501
1188929389 M * daniel_hozac 2.4? whatever is using the ioctl-based interfaces to get/set flags/xid/etc.
1188929406 M * micah hmm, i guess debian doesn't need to even enable legacy
1188929407 M * Bertl thought so, especially micah's kernel does not need it at all
1188929425 M * Bertl so maybe an even better approach would be to disable that
1188929458 M * micah disabling means passing --enable-apis=NOLEGACY?
1188929475 M * daniel_hozac no, disabling means accepting the default :)
1188929501 M * micah ah, yes I see I have --enable-apis=NOLEGACY set, so I can just remove that from the configure flags
1188929514 M * daniel_hozac right.
1188929518 M * matti Hi Bertl, daniel_hozac 
1188929523 M * daniel_hozac hi matti
1188929662 M * Hollow hey matti
1188929908 M * micah looks like it builds fine w/o patch and w/o --enable-apis=NOLEGACY
1188930101 M * micah ok, new util-vserver uploaded with these changes (including yours cryptronic)
1188930122 M * cryptronic thanks :)
1188930330 M * MooingLemur does 2.3.0.17 include the IPv6 patch?
1188930384 M * Bertl yep, most parts of that
1188930432 M * MooingLemur nice :)
1188930533 M * bXi i think i found a bug
1188930536 M * daniel_hozac i'd say more parts of that.
1188930546 M * bXi daniel_hozac: your the network dude right?
1188930556 M * daniel_hozac Bertl's the kernel dude :)
1188930566 M * bXi well
1188930572 M * bXi i have a guest with ipv6 only
1188930586 M * bXi i started irssi
1188930592 M * bXi connected to irc6.oftc.net
1188930602 M * bXi it resolvs an ipv4 address and connects to that
1188930623 M * MooingLemur then space-time formed a rift?
1188930639 M * bXi something like that
1188930651 M * daniel_hozac micah: i guess the backport will require the patch and switch though.
1188930656 M * Bertl bXi: let's wait for 2.3.0.18 (should be done tonight)
1188930679 M * bXi Bertl: in how many hours is that? (seeing as i'm in a different timezone?)\
1188930724 M * nanonyme hmm, i think irssi prefers A records unless it's set to prefer AAAA
1188930751 M * nanonyme which is silly :p
1188930755 M * micah daniel_hozac: yeah
1188930757 M * Bertl bXi: I'd say 4-7 hours
1188930771 M * bXi nanonyme: correct but my vserver with ONLY a ipv6 address shouldnt be able to see ipv4 ip's
1188930780 M * Bertl correct
1188930780 M * nanonyme ah
1188930801 M * Bertl bXi: unless you have lback stuff enabled
1188930811 M * Bertl bXi: in which case, you already have ipv4 too
1188930822 M * bXi shouldnt i see that in ifconfig or something then?
1188930829 M * Bertl yep
1188930837 M * fb bXi: it's not true
1188930856 M * bXi my ifconfig only states the ipv6 address
1188930866 M * fb bXi: there's nothing wrong that DNS server returns A record, instead AAAA
1188930875 M * bXi thats not wrong
1188930887 M * bXi but it shouldnt be able to connect from there
1188930893 M * fb in ipv4-only env you can ask for AAAA records
1188930913 M * ard but in an ipv6 only env you can't connect to ipv4 servers
1188930923 M * fb exactly
1188930932 M * ard that's what bXi said...
1188930968 M * fb ard but your client can ask for ipv4 address (A record) in ipv6-only env
1188930982 M * bXi yes correct
1188930984 M * ard that's ok, but in bXi's case it connected with ipv4
1188930988 M * bXi but shouldnt be able to connect to one
1188930992 M * ard which it should not have
1188931004 M * fb ok, i must have misunderstood
1188931022 M * daniel_hozac i think this is actually an ancient bug.
1188931049 M * daniel_hozac for a rather long time, a guest with no IPv4 address has been able to connect to things.
1188931064 M * bXi hmmm
1188931068 M * bXi have you fixed it?
1188931076 M * fb daniel_hozac: btw, with 2.2.0.x i can use network without any interface configured. is that normal?
1188931083 M * daniel_hozac i'm not sure whether it's still the case.
1188931085 M * bXi (i'm running 2.3.0.17)
1188931107 M * daniel_hozac fb: which is what i just said :)
1188931112 M * ard :-)
1188931126 M * fb :-)
1188931161 M * bXi daniel_hozac: i'm loving that proper ipv6 support btw
1188931168 M * fb daniel_hozac: so with latest 2.2 it is
1188931177 A * ard actually uses 2.2.0 and openbsd-inetd with sshd -i to be able to connect with ipv6
1188931201 M * ard that's incoming only of course...
1188931204 M * fb ard: there's quite stable ipv6 patch for 2.2
1188931228 M * daniel_hozac fb: i'd actually call 2.3 more stable with regard to IPv6.
1188931230 M * ard yes... I've tried it this morning for the first time. Due to that I was an hour late on my work
1188931234 M * fb and there's grsec version too
1188931254 M * fb daniel_hozac: ok, it works charmless for some time for me ;)
1188931261 M * fb harmless*
1188931361 J * bluepunk root@void.apparoat.nl
1188931373 M * bluepunk yes yes i know dont irc as root :P
1188931415 M * nanonyme and if you irc with root, keep identd off
1188931439 M * bXi hehe
1188931443 M * bXi bluepunk == me
1188931453 M * nanonyme kk
1188931474 M * bXi void.apparoat.nl is one of my play vservers
1188931483 M * nanonyme some irc servers block root irc with identd on
1188931572 M * sid3windr and because of that you advise to disable ident :p
1188931767 Q * bluepunk Quit: leaving
1188931823 M * nanonyme sid3windr, exactly. hackers don't know if it's real root if you have identd off
1188931939 M * daniel_hozac they don't know if it's real root if you identd on...
1188931977 M * sid3windr indeed
1188931978 M * bXi venice ~ # vserver void die
1188931981 M * fb unless you have outgoing connection on privileged port
1188931982 M * sid3windr I don't really see the difference
1188931985 M * bXi daniel_hozac: you need to implement that :P
1188931989 M * sid3windr it's the one account you know to be there :p
1188931998 M * daniel_hozac bXi: how would it differ from stop?
1188932008 M * bXi daniel_hozac: not
1188932020 M * bXi just for use when your frustrated about a certain guest again
1188932027 M * sid3windr how about vserver void die die die die die die die die 
1188932027 M * bXi maybe delete it could be usefull :P
1188932032 M * sid3windr some kind of --force ;)
1188932034 M * bXi sid3windr: that would do yes :P
1188932080 J * bluepunk ~root@ut2004.apparoat.nl
1188932282 M * bXi RX bytes:294940128194 (281276.8 Mb)  TX bytes:3612329183 (3444.9 Mb)
1188932287 M * bXi i wonder what this vserver has been doing
1188932337 M * daniel_hozac those stats are for the entire host.
1188932348 M * daniel_hozac or rather, the interface.
1188932354 M * bluepunk still
1188932379 M * bluepunk thats quite a lot for a box which has been installed around 2 months ago (i think around when 2.3.0.17 came out
1188932418 Q * bluepunk Quit: leaving
1188932429 J * blupeunk ~root@cl-998.ams-05.nl.sixxs.net
1188932461 M * Bertl well, depends, 300GB a month is roughly 100k/s or so
1188932501 M * bonbons Bertl: did you see my mention from yesterday evening regarding guest's init PID not always being show as 1 to guest?
1188932531 M * Bertl bonbons: yep, daniel_hozac has a test fix for that
1188932550 M * Bertl (which isn't quite mature yet, IIRC)
1188932572 M * Bertl daniel_hozac: btw, what did the test without your patch give?
1188932680 M * bonbons ok, was looking at your patch-collection and didn't see anything :), will look at daniel's
1188932907 M * bonbons went to bed before anything got out... bound to BTZ which is currently between CEST and UTC...
1188932909 M * bonbons the part related to unix sockets is not covered yet by the patches
1188932980 N * virtuoso_ virtuoso
1188933042 Q * blupeunk Quit: Changing server
1188933664 J * igraltist ~user4@kasten-edv.de
1188933712 J * prae ~benjamin@mon75-17-88-175-211-122.fbx.proxad.net
1188933856 Q * brcc_ Ping timeout: 480 seconds
1188934359 Q * prae Remote host closed the connection
1188934371 J * prae ~benjamin@mon75-17-88-175-211-122.fbx.proxad.net
1188934386 J * root____1 ~root@cl-998.ams-05.nl.sixxs.net
1188934402 M * root____1 test
1188934415 N * root____1 bluepunk
1188934697 M * bluepunk this irssi proxy stuff is kinda nice
1188934860 M * daniel_hozac Bertl: ah, i got distracted.
1188934885 Q * prae Quit: Pwet
1188935059 M * daniel_hozac Bertl: hmm, seems to happen even without the patch.
1188935083 M * daniel_hozac i'm thinking COW, as i only made these guests quick clones.
1188935111 J * xinming ~xinming@125.109.243.31
1188935116 M * xinming anyone here uses gentoo?
1188935135 M * daniel_hozac i'd hope so...
1188935151 M * xinming Ok, just a question.
1188935157 M * Bertl daniel_hozac: interesting
1188935215 J * coderanger_ ~coderange@wireless-19-150.media.mit.edu
1188935699 M * xinming anyone run a X within vserver?
1188935725 M * Bertl xinming: yep, a few folks do, both server and client
1188935736 M * xinming I didn't try. But will do soon... :-)
1188935779 M * xinming If we need maintaince on host, Just ssh to the host, and su. :-)
1188936360 Q * sladen Ping timeout: 480 seconds
1188936704 J * sladen paul@starsky.19inch.net
1188937825 Q * bluepunk Quit: leaving
1188938203 Q * dna Quit: Verlassend
1188938339 Q * xinming Quit: leaving
1188938369 M * daniel_hozac Bertl: got it.
1188938666 Q * meandtheshell Quit: Leaving.
1188938698 M * daniel_hozac http://people.linux-vserver.org/~dhozac/p/k/delta-cow-fix12.diff
1188938745 M * daniel_hozac i'm rebooting to test it now, but the debugging output showed -2 returned from path_lookup(old)
1188938815 M * daniel_hozac (i had forgotten to reset the xid tagging when setting IMMUTABLE|IUNLINK)
1188938834 M * daniel_hozac -xid :)
1188939103 M * daniel_hozac yep, all guests stopped nicely now...
1188939177 Q * bonbons Quit: Leaving
1188939207 M * daniel_hozac the other issue bonbons was talking about was that SCM_CREDENTIALS doesn't have initpid virtualization.
1188939488 J * yarihm ~yarihm@84-75-130-73.dclient.hispeed.ch
1188939911 Q * cryptronic Quit: Leaving
1188940105 J * Aiken ~james@ppp121-45-255-55.lns2.bne4.internode.on.net
1188942250 N * DoberMann DoberMann[ZZZzzz]
1188942966 Q * onox Quit: zZzZ
1188944174 M * Bertl daniel_hozac: excellent!
1188944195 M * Bertl daniel_hozac: yeah, I think that needs some work too
1188944350 Q * Piet Ping timeout: 480 seconds
1188944472 J * Piet hiddenserv@tor.noreply.org
1188946026 Q * yarihm Quit: Leaving
1188948194 Q * Piet Quit: Piet