1140135013 M * pflanze Is it safe to use the UML 'SKAS' patch with vserver?
1140135094 M * Bertl pflanze: probably not :)
1140135146 J * grant mep@p5091AE01.dip0.t-ipconnect.de
1140135154 M * Bertl wb grant!
1140135175 P * meandtheshell 
1140135565 Q * grant_ Ping timeout: 480 seconds
1140135874 P * stefani I'm Parting (the water)
1140136003 Q * cdrx Quit: Leaving
1140136712 Q * matt1 Quit: Download Gaim: http://gaim.sourceforge.net/
1140136848 J * matta ~matta@c-68-32-239-173.hsd1.pa.comcast.net
1140136850 P * matta 
1140136861 M * pflanze Am I a wimp if I wish a iptables debugger?
1140136991 M * Bertl heh, there is a test mode, i.e. you can trace a test packet ...
1140137066 M * pflanze hm. so would you insert counting rules at strategic points or so? like printf statements?
1140137119 A * pflanze thinks to have found the rules bug in the meantime
1140137674 A * pflanze did error, still not found
1140138154 M * pflanze Bertl: where do I find info about those test means?
1140138207 M * pflanze any google keyword tips?
1140138437 J * rs ~rs@vol75-7-82-229-177-124.fbx.proxad.net
1140138445 M * rs hi there
1140138453 M * Bertl hi rs! :)
1140138457 M * Bertl pflanze: http://iptables-tutorial.frozentux.net/scripts/rc.test-iptables.txt
1140138461 M * rs how are you bertl ?
1140138487 M * Bertl thanks, fine, and you?
1140138519 M * rs fine fine, a lot of work but fine :)
1140139669 J * matta ~matta@71.224.125.126
1140139771 M * Bertl wb matta!
1140140047 M * pflanze Thanks for the link, Bertl.
1140140094 M * pflanze Since this machine is being used live, so too many packets would be shown, and since ping works for me (only tcp connections not), I leave that for another day.
1140140400 J * F00x ~Staff@tor-irc.dnsbl.oftc.net
1140140629 M * Bertl welcome F00x!
1140140761 M * F00x thx friends
1140140762 M * F00x :)
1140140794 M * F00x good night! to all ;)
1140140801 M * F00x nite Bertl
1140140803 M * F00x :)
1140140807 Q * F00x Quit: 
1140141005 Q * gerrit Ping timeout: 480 seconds
1140141046 J * Method Method@c-69-243-91-43.hsd1.md.comcast.net
1140141088 M * pflanze Hm, with older vserver / kernel 2.4 'wall' or 'shutdown' would send messages to all vservers as well.
1140141091 M * Bertl welcome Method!
1140141106 M * pflanze Now with 2.0.1/2.6 it only sends to users logged on to the host.
1140141112 M * Method hi
1140141131 M * pflanze I'd like to notify vserver users as well. Am I missing something?
1140141160 M * Bertl pflanze: strace -fF of the wall?
1140141193 J * Aiken_ ~james@tooax6-157.dialup.optusnet.com.au
1140141212 M * pflanze Bertl: what are you looking for exactly? Or do you want me to send you the strace?
1140141225 M * Bertl upload it somewhere please
1140141326 M * pflanze Bertl: http://elvis-jaeger.mine.nu:5080/~chris/scratch/wall.strout
1140141394 M * Bertl well, looks to me like it is sending to each /dev/pts
1140141409 M * pflanze from what I can say, it opens /var/run/utmp and slurps it in.
1140141410 M * Bertl as the pts are isolated, it won't see those of the guests
1140141526 M * Bertl running wall in xid=1 might help there
1140141531 M * pflanze nope, chcontext --ctx 1 doesn't help
1140141550 Q * Aiken Ping timeout: 480 seconds
1140141552 M * Bertl then you need a better wall
1140141553 M * pflanze then it is not written to any context anymore, not even host.
1140141598 M * pflanze chcontext --ctx 1 ls /dev/pts/ doesn't show any contents :)
1140141658 M * pflanze hmm,
1140141658 M * pflanze chcontext --ctx 1005 ls /vservers/5/dev/pts/
1140141658 M * pflanze No such file or directory
1140141675 M * Bertl namespaces?
1140141680 M * pflanze yes.
1140141700 M * Bertl well, that was a rhetoric question :)
1140141717 M * pflanze Mine was a rhetoric answer ;)
1140141725 M * Bertl k
1140141749 M * Bertl well, I'm inclined to consider the /dev/pts in xid=1 a bug
1140141752 M * pflanze (since I didn't knew anyway for sure  but anyway.)
1140141855 M * Bertl what kernel/patch version do you have
1140142055 M * pflanze currently 2.6.14.4-vs201
1140142059 M * pflanze going to reboot now.
1140142139 M * Bertl if you want to test, I can suggest a simple modification
1140142305 M * pflanze the wall thing?
1140142335 M * pflanze Not so important right now. And wouldn't the right way to be to enter all namespaces? something for a good vserver library..
1140142402 A * pflanze is now running 2.6.15.4-vs2.0.2-rc5
1140142639 M * Bertl yep, but maybe pts visibility in xid=1 would be sufficient
1140142752 J * matt1 ~matta@71.224.125.126
1140142895 Q * Method Ping timeout: 480 seconds
1140142993 Q * Aiken_ Remote host closed the connection
1140143210 Q * matta Ping timeout: 480 seconds
1140143219 J * matta ~matta@71.224.125.126
1140143417 M * pflanze Somehow my box has (now?) difficulties starting some services upon boot, very strange. The host doesn't start ntpd, and my proxy vserver doesn't start httpd.
1140143424 M * pflanze nothing in the logs.
1140143441 A * pflanze going to reboot again to see again
1140143475 Q * matt1 Ping timeout: 480 seconds
1140143490 A * pflanze notices that even before the last reboot the problem was there. so  it's not vserver version related.
1140143517 M * pflanze (but it is new)
1140143646 Q * entroposcope Quit: Ping timeout
1140143760 J * Aiken ~james@tooax6-157.dialup.optusnet.com.au
1140143910 M * Skram where is Superklag?
1140144071 J * Method Method@c-69-243-91-43.hsd1.md.comcast.net
1140144214 M * pflanze ahh,
1140144238 M * pflanze those services require dns running, and the dns vserver has not finished booting at that time.
1140144241 M * pflanze (parallel startup)
1140144415 J * ntrs ~ntrs@68-188-37-15.dhcp.stls.mo.charter.com
1140144840 Q * matta Ping timeout: 480 seconds
1140145520 Q * Doener Quit: Leaving
1140145741 M * pflanze Hmmm, can it be that this socket separation stuff has went into vserver only after patch-2.6.14.3-vs2.01.diff.gz ?
1140145765 M * pflanze Since: what I did before and worked, does not anymore now with 2.6.15.4-vs2.0.2-rc5, Bertl.
1140145772 M * pflanze (the uml networking stuff)
1140145799 M * pflanze uml_switch says: "send_sock sending to fd 4 Connection refused"
1140145806 M * Bertl well, let's check (probably I can dig out that version :)
1140145878 M * Bertl http://vserver.13thfloor.at/Devel/PAT-2.0.2/delta-usock-fix01.diff
1140145922 M * pflanze (from strace:  sendto(4, "\377\377\377\377\377\377J\314v\t\f\332\10\6\0\1\10\0\6"..., 42, 0, {sa_family=AF_FILE, path=@}, 110) = -1 ECONNREFUSED (Connection refused) )
1140145942 M * Bertl that's something different
1140145957 M * Bertl refused means, the other side rejected it
1140146314 M * pflanze (yes, the delta-usock patch is in the new kernel I'm using, but not the old)
1140149690 J * hallyn_ ~xa@c-24-11-243-196.hsd1.in.comcast.net
1140149690 Q * hallyn Read error: Connection reset by peer
1140150238 M * Bertl okay, off to bed now .. back tomorrow!
1140150247 M * Bertl have a good one, everyone! cya!
1140150254 N * Bertl Bertl_zZ
1140150255 M * rs good night bertl
1140151206 J * gerrit ~gerrit@c-67-160-130-59.hsd1.wa.comcast.net
1140152070 Q * mire Ping timeout: 480 seconds
1140152833 J * matta ~matta@71.224.125.126
1140152843 Q * matta Quit: 
1140154477 J * Viper0482 ~Viper0482@p54977FBA.dip.t-dialin.net
1140155864 Q * pflanze Quit: good night
1140156228 J * _nokoya young@hi-230-82.tm.net.org.my
1140156228 Q * nokoya Read error: Connection reset by peer
1140156234 N * _nokoya nokoya
1140156606 Q * Viper0482 Quit: bin raus, 
1140157917 J * tudenbart ~willi@xdsl-213-196-250-223.netcologne.de
1140158288 Q * dothebart Read error: Connection reset by peer
1140160358 J * mire ~mire@183-166-222-85.COOL.ADSL.VLine.verat.net
1140162995 Q * Smutje Ping timeout: 480 seconds
1140163789 Q * andrew_ Quit: leaving
1140164001 J * Smutje ~Smutje@xdsl-87-78-41-125.netcologne.de
1140165164 Q * mkhl Quit: 
1140165765 Q * shedi Quit: Leaving
1140168283 Q * rs Quit: rs
1140168775 J * prae ~prae@ezoffice.mandriva.com
1140169991 J * lilalinux ~plasma@h1-gw.of.net-lab.net
1140171825 J * meandtheshell ~markus@85-124-10-40.dynamic.xdsl-line.inode.at
1140172239 Q * alexx Read error: Connection reset by peer
1140172622 M * ebiederm_oO dhansen:  I have signed up to present a paper "Multiple Instances of the Global Linux Namespaces" at OLS so I better make it :) 	
1140173240 N * ebiederm_oO ebiederm_zZ
1140175111 J * oliwel ~oliwel@ldvpc07.ldv.e-technik.tu-muenchen.de
1140175235 M * oliwel Hi
1140175497 A * oliwel waves hello to the crowd
1140175501 M * oliwel anybody in
1140175608 M * oliwel Hollow: ping ?
1140175698 M * Hollow oliwel: pong
1140175705 M * oliwel ahh :)
1140175721 M * oliwel Can you pls point me to the URL where I can download the fixes for the kernel bugs
1140175733 M * Hollow which kernel bugs?
1140175744 M * oliwel I recompiled a machine and I urgently ned to fix the isue with the mountpoints I have...
1140175754 M * oliwel You remeber my Problem with blocked mount points ?
1140175774 M * Hollow ah, http://vserver.13thfloor.at/Experimental
1140175808 M * oliwel yeah this one - you know whcich subdir contains the patches for the bugs ? I am searching for a while and cant dinf it
1140175948 M * oliwel hmm useless - i cant remember the name..
1140175966 M * oliwel can you help me out ;)
1140176001 M * Hollow hm
1140176053 M * Hollow hm, the delta is not there... maybe this one? http://vserver.13thfloor.at/Devel/PAT-2.1.1/delta-bme-fix01.diff
1140176336 M * oliwel no :(
1140176352 M * oliwel I will remember when I see the oatch - it was someting in the network.h...
1140176362 M * oliwel is there a way to search the ICQ Logs ?
1140176366 M * oliwel IRC Logs of course...
1140176390 M * Hollow http://irc.13thfloor.at/LOG/
1140176501 M * oliwel ok thx - will grqab and search it locally
1140176612 M * oliwel http://vserver.13thfloor.at/Devel/PAT-2.0.1/delta-indev-fix01.diff :)
1140176707 M * oliwel F++k - the patch is already in the new release - but the problem persists...
1140176749 M * RoadRunnR oliwel: i had the same problem and the patched fixed it for me
1140176770 Q * Aiken Ping timeout: 480 seconds
1140176782 J * yodahome ~yoda@pD9515199.dip0.t-ipconnect.de
1140176784 N * Bertl_zZ Bertl
1140176787 M * Bertl morning folks!
1140176789 M * oliwel Seems not to work in all circumstances - I exactly have the same issues than before...but the patch is alreay in the 2.1.0.4...
1140176791 M * oliwel Hi Bertl 
1140176803 M * Bertl what issues?
1140176807 M * oliwel I have bad news - probably
1140176817 M * yodahome Hi @ all!
1140176824 M * oliwel You remember my problem with blocked mountpoints ?
1140176835 M * Bertl yup?
1140176845 M * oliwel that we assumed fixed with the delta-indev patch 
1140176853 M * oliwel seems that it is back
1140176864 M * oliwel just setup a new machine this week
1140176873 M * yodahome anybody here ever made an ubuntu vserver on a fedora core 4 system? (successfully)
1140176875 M * oliwel just have one vserver guest running (2.1.0.4)
1140176887 M * oliwel but cant umount a drbd device even with the guest shut down
1140176921 M * Bertl let me check if 2.1.0.4 does even contain the fix
1140176975 M * oliwel it does
1140177001 M * oliwel i dl'ed the patch because it was my first thought - but it echos a "already applied"
1140177153 M * Bertl it's not the indev fix
1140177159 M * oliwel It might be an issue only with "dying" guests - after the migration I had one failed bootup of the guest - might be this breaks it ?
1140177166 M * oliwel ok - what then ?
1140177188 M * oliwel I searched the IRC Logs and this was the propsed and working solution
1140177206 M * Bertl http://vserver.13thfloor.at/Experimental/delta-rawsend-fix01.diff
1140177248 M * Bertl and another one, sec
1140177252 M * oliwel ok
1140177303 M * Bertl http://vserver.13thfloor.at/Devel/PAT-2.0.2/delta-proc-fix02.diff
1140177314 M * Bertl I'd suggest to use one of the -rc*
1140177333 M * Bertl this will give you a working version, I guess
1140177356 M * Bertl 2.0.2-rc* or 2.1.1-rc*
1140177371 M * oliwel but they are against 2.6.16 arent they ?
1140177543 M * oliwel this one patch-2.6.16-rc1-vs2.1.0.6.1.diff  
1140177557 M * Bertl 2.6.15.4 and 2.6.16, yes
1140177630 M * oliwel as gentoo has not released .16 - I will try this patch-2.6.15.4-vs2.1.1-rc6.diff    - might taht be suitable ?
1140177643 M * oliwel Hollow: or is there an ebuild in preparation ?
1140177652 M * Bertl there should be an up-to-date gentoo release for them, IIRC
1140177702 M * Hollow no
1140177708 M * Hollow no 2.6.16 yet
1140177714 M * Bertl but 2.6.15.4 or?
1140177724 M * Hollow is there a patch for it?
1140177731 M * Bertl sure, plenty of
1140177750 M * Hollow hm, thought the rc patches are againt 2.6.16 only
1140177765 M * Bertl nope, both mainline branches
1140177814 J * shedi ~siggi@tolvudeild-196.lhi.is
1140177847 M * Hollow will make an ebuild shortly
1140177873 M * oliwel Hollow: would be great - put in the patches by hand and rebooting atm...
1140177930 M * Bertl oliwel: hmm, the two fixes or the 2.6.15.4 rc?
1140178020 Q * mcp Ping timeout: 480 seconds
1140178062 M * oliwel yes
1140178078 M * oliwel applied and machine is back alive..
1140178089 M * oliwel but my proc/mount has gone inside the guest... :(
1140178165 M * oliwel Bertl: any problems with the patches against 15.4 ?
1140178185 M * oliwel its a produciton system - so it should be reliabe
1140178358 M * oliwel problem persists...
1140178475 M * oliwel Next issue: my etc is on a read-only filesystem and due to this its a symlink to /var - the file is not updated correctly...only the last line appears in the file - former solution to link to proc/mounts i ineffectife becaous proc/mount is gone...
1140178507 M * oliwel I will reboot into the old kernel - this seems to be bit fuzzi..
1140178613 M * oliwel Hollow: will the patches apply against gentoo-sources ?
1140178651 M * Hollow it will fail in sysctl.h but it's easy to fix
1140178674 M * oliwel ok - I will try
1140178737 M * oliwel the Makefile fails also
1140178748 M * Hollow yeah, well... that's only the versio
1140178749 M * Hollow n
1140178872 M * oliwel kk
1140178978 M * oliwel New kernel is on the way :)
1140178987 M * oliwel hope it will reboot ...
1140179005 M * oliwel Bertl: Any ideas on the persisting problem ?
1140179043 M * Bertl what problem?
1140179107 M * Bertl oliwel: if you encounter issues, please try with a recent mainline kernel/patch first
1140179165 M * oliwel Bertl: the mountpoint is still blocked and my proc/mount has gone inside the guest
1140179205 M * oliwel Bertl: I am just compiling a fresh kernel with the recent patch
1140179248 M * Bertl that's a good idea
1140179632 J * mcp ~hightower@wolk-project.de
1140179816 M * Bertl wb mcp!
1140180180 M * oliwel Bertl:  Problem remains
1140180190 M * oliwel  2.6.15-gentoo-r4-vs2.1.1-rc6
1140180250 M * oliwel Bertl: I have to leave - will write a short summary to the list...
1140180386 M * Bertl k, but I somewhat doubt that result ...
1140180410 M * oliwel I cant tell you anything else...
1140180535 M * Bertl okay, will verify with 2.6.15.4-vs2.1.1-rc*
1140180542 M * oliwel mail is on the way - have to go...
1140180549 M * Bertl cya, and tx!
1140180560 M * oliwel cu and thx for your time...
1140180567 M * oliwel will be back in h
1140180571 M * oliwel approx 2 h
1140180588 Q * oliwel Quit: Chatzilla 0.9.69.1 [Firefox 1.5.0.1/2006021510]
1140181161 J * nayco ~nayco@lns-bzn-29-82-64-225-98.adsl.proxad.net
1140181216 M * nayco hello, all !!
1140181377 M * Bertl welcome nayco!
1140181720 J * rs ~rs@vol75-7-82-229-177-124.fbx.proxad.net
1140181769 M * Bertl welcome rs!
1140181796 M * rs hi
1140181822 M * nayco Bertl: I goona buy two new servers : One with an single Opteron, and the other with two of them ;-)
1140181835 M * nayco Of course, the kernel will be patched...
1140181885 M * Bertl cool!
1140181928 M * nayco So, I think this is gonna be a nice occasion to see how the vserver perform on it...
1140181991 M * nayco Have you had many feedback of linux-VServer on (dual)opteron ?
1140182017 M * Bertl yes, they seem to work quite well
1140182237 M * nayco Ok, great.
1140183600 Q * Method Read error: Connection reset by peer
1140184515 Q * rs Quit: rs
1140185093 J * oliwel ~mail-at-o@host-62-245-151-178.customer.m-online.net
1140185348 J * robig ~robig@154.250.ddkom.net
1140185355 M * robig hey guys
1140185375 M * robig Hi Bertl! :)
1140185392 M * robig i've a little problem again
1140185392 M * phreak`` oliwel: just do an emerge --sync in about half an hour :)
1140185427 M * phreak`` morning folks (to lend Bertl's quote) :)
1140185458 M * robig ive installed exim on a vserver, but when starting it... I got address already in use IP
1140185495 M * robig i guess another vserver is usig IP:25 already, how can i find out which vserver it is?
1140185517 M * phreak`` robig: ensure that the host itself doesn't run exim on 0.0.0.0:25 ;)
1140185562 M * robig ive shut it down with /etc/init.d/exim4 stop on the host
1140185593 M * robig host:~# lsof |grep smtp
1140185600 M * robig outputs nothing
1140186013 M * robig how do u preffer to configure multiple vservers mailing to the outside?
1140186018 M * robig -f
1140186244 M * oliwel Bertl: Hi - so I am back
1140186255 M * oliwel did you make any progress 
1140186421 J * Arezzo2006^^ ~e250382@host71-210.pool80116.interbusiness.it
1140186438 M * Arezzo2006^^ @help
1140186442 M * Bertl welcome Arezzo2006^^!
1140186451 M * Bertl oliwel: nope, seems to work fine here!
1140186582 P * Arezzo2006^^ 
1140186802 Q * yodahome Ping timeout: 480 seconds
1140187039 M * oliwel Bertl: with DRBD ?
1140187071 M * Bertl no, no drbd here, but the devices mounted get unlocked
1140187086 M * Bertl there is a tool (and a simple patch) to check for holders
1140187089 M * oliwel hmm - so might be a problem with drbd...I dont know how it detects if its mounted
1140187098 M * oliwel can you send me this tool ?
1140187103 M * oliwel might be helpful
1140187103 M * Bertl http://vserver.13thfloor.at/Experimental/patch-holders.diff
1140187106 M * oliwel ;)
1140187107 M * Bertl http://vserver.13thfloor.at/Experimental/holders-0.02.tar.bz2
1140187122 M * oliwel perhaps this shows me the source of the probs
1140187172 M * Bertl you should definitely try with mainline + patch + drbd
1140187188 M * Bertl and if that fails too, just with mainline + patch
1140187241 M * oliwel Should do...shit..plenty of work to do
1140187257 M * oliwel so you use vanilla and standard block devs ?
1140187502 M * oliwel Bertl: One last Question - is it intended that /proc/mounts disappears inside a guest ?
1140187557 M * Bertl no, definitely not
1140187613 M * Bertl chcontext --xid 42 -- wc /proc/mounts 
1140187613 M * Bertl New security context is 42 4      24     119 /proc/mounts
1140187646 M * Bertl that's with 2.6.15.4-vs2.1.1-rc*
1140187750 M * oliwel So it seems that somewhat here is really broken...
1140187781 M * oliwel I hope that it is not an drbd issue
1140188519 M * Bertl well, could be, who did combine them?
1140188534 M * Bertl (i.e. maybe there are conflicts between the patches)
1140188999 P * robig 
1140189005 Q * FireEgl Read error: Connection reset by peer
1140189066 M * locksy the addition of the #include <linux/mount.h> to include/linux/fs.h in 2.0.2-rc5 breaks the build of drivers/mtd/devices/blkmtd.c due to duplicate (conflicting) definitions of name_to_dev_t
1140189147 M * locksy it's a simple patch to fix it but I haven't bothered (I just disabled CONFIG_MTD_BLKMTD instead)
1140189194 J * matta ~matta@c-68-32-239-173.hsd1.pa.comcast.net
1140189198 M * Bertl well, good idea, or do you have MTD devices?
1140189220 M * Bertl locksy: but it was already noticed and will be fixed in the next release
1140189244 M * locksy Oh, well :) 
1140189244 J * FireEgl Atlantica@Atlantica.Tcldrop.Org
1140189253 M * Bertl locksy: but thanks for the input!
1140189256 M * locksy Two bug reports are better then none...
1140189268 M * Bertl exactly my words
1140189694 J * rs ~rs@office.dailymotion.com
1140190819 Q * vrwttnmtu Quit: Leaving
1140192390 J * Viper0482 ~Viper0482@p54977FBA.dip.t-dialin.net
1140192396 M * oliwel Bertl:There ?
1140192733 Q * eyck Quit: leaving
1140192865 Q * shedi Quit: Leaving
1140192880 M * Bertl oliwel: yup!
1140193181 J * liquid3649_ ~Viper0482@p54977FBA.dip.t-dialin.net
1140193223 M * shuri hi there
1140193254 Q * liquid3649_ Quit: 
1140193472 Q * Viper0482 Ping timeout: 480 seconds
1140193548 J * eyck eyck@81.219.64.71
1140194370 Q * DuckMaster Ping timeout: 480 seconds
1140194370 Q * Duckx Ping timeout: 480 seconds
1140194591 J * DuckMaster ~duckx@195.75.27.158
1140194618 J * Duckx ~duckx@195.75.27.158
1140194825 M * oliwel Bertl: FYI - currently running vanilla 2.6.15.4 + drbd + vserver - problems persist: no /proc/mounts and drbd device blocked after guest shutdown
1140194907 M * Bertl okay, next step vanilla+ vserver?
1140194919 M * oliwel Yes, compling atm
1140194929 M * oliwel hw does the holder tool work ?
1140194951 M * oliwel ./holder /dev/drbd/wwwgc
1140194953 M * oliwel ioctl:: Device or resource busy
1140194958 M * oliwel any more output possible ?
1140195035 M * Bertl check dmesg
1140195041 M * Bertl (reuires the patch)
1140195045 M * Bertl *requires
1140195118 M * oliwel bdev[c20e8480]: holder=c03b99a0[2]
1140195124 M * oliwel what does this tell me ?
1140195141 M * Bertl that two holders exist for that device
1140195181 M * Bertl when you start, it will show 0 holders, I presume
1140195300 M * oliwel might that be the underlying md device ?
1140195316 M * oliwel I get the same output when I access /dev/mdX
1140195332 M * Bertl well, holders are _above_ the device
1140195344 M * Bertl i.e. they are keeping a reference to that device
1140195398 M * oliwel hm..
1140195417 M * oliwel is there a way to find out "what" the handler is
1140195526 M * oliwel hmm - seems beside that drbd tools keep the handlers - even those devices that are unused have handlers
1140195579 M * Bertl that looks at least strange ...
1140195589 M * oliwel yup
1140195608 J * fwl ~fwl@83.215.237.1
1140195611 M * Bertl is the device deactivated
1140195617 M * Bertl welcome fwl!
1140195627 M * fwl hi Bertl
1140195659 M * oliwel the drbd is shutdown, the corresponding "worker" process died...
1140195690 M * oliwel booting vanilla-vs - lets see whats up here...
1140196116 M * Hollow oliwel: 2.0.2-rc5 and 2.1.1-rc6 should be in portage now..
1140196129 M * Hollow phreak`` put them in..
1140196140 J * Viper0482 ~Viper0482@p54977FBA.dip.t-dialin.net
1140196222 Q * prae Quit: Execute Order 69 !
1140196340 M * oliwel Hollow: thx - currently using vanilla :)
1140196368 M * Bertl okay, off for dinner now ... back shortly
1140196377 N * Bertl Bertl_oO
1140196536 M * oliwel Bertl_oO: :(
1140196711 J * S2k Legalize@dial33-210.brunet.bn
1140196938 P * S2k 
1140196997 Q * fwl Quit: This computer has gone to sleep
1140197071 M * oliwel Bertl_oO: The problem persists - vanilla + vs with a standard block device....pls ping when back from dinner
1140197529 N * Bertl_oO Bertl
1140197540 M * Bertl oliwel: ping
1140197570 M * Bertl oliwel: uname -a
1140197617 M * oliwel Hi Bertl - enjoyed your meal ;)
1140197626 M * Bertl yeah, was quite good ...
1140197643 M * oliwel *g*
1140197645 M * oliwel 2.6.15.4-vs2.1.1-rc6
1140197681 M * Bertl okay, give me a minute to check the code once again
1140197729 M * Bertl ah, and please check on your side if lsof +d  or lsof +D shows something for that device?
1140197745 M * Bertl also make sure that you have no automounter or nfsd active
1140197828 M * oliwel I dont have automount or nfs running - the device is not mounted according to /proc/monut
1140197844 M * Bertl any guests are running?
1140197846 M * oliwel BUT: I stll dont have /proc/mounts in my guests
1140197850 M * oliwel no the guest is down
1140197891 M * Bertl okay, I need the complete testme.sh output for a start
1140197905 M * oliwel mom
1140197911 M * Bertl (please get v0.15 and upload the output somewhere)
1140197934 M * oliwel can you give me the url
1140197947 M * Bertl http://vserver.13thfloor.at/Stuff/SCRIPT/testme.sh-0.15
1140197995 M * oliwel http://pastebin.com/559825
1140198166 M * Bertl what does 'chcontext --xid 42 --secure -- wc /proc/mounts' give you?
1140198213 M * Bertl what gcc do you use?
1140198244 M * oliwel New security context is 42
1140198245 M * oliwel wc: /proc/mounts: No such file or directory
1140198254 M * oliwel gcc version 3.4.5 (Gentoo 3.4.5, ssp-3.4.5-1.0, pie-8.7.9)
1140198298 M * Bertl $ chcontext --xid 42 --secure -- wc /proc/mounts
1140198298 M * Bertl New security context is 42 4      24     119 /proc/mounts
1140198315 M * Bertl so I have _no_ idea why this doesn't work for you ...
1140198351 M * Bertl do you get any messages in tge kernel log?
1140198383 M * oliwel possible that procunhide wasnt execeuted ?
1140198400 M * oliwel dynamic contexts disabled. :)
1140198438 M * oliwel but regarding the "real" servers I have given them a fixed contextid...
1140198439 M * Bertl vprocunhide should not affect /proc/mounts
1140198448 M * oliwel Any other ideas ?
1140198454 M * Bertl /proc/mounts is a symlink
1140198462 M * Bertl it points to self/mounts
1140198490 M * Bertl so this basically means that the pid info is broken
1140198500 M * Bertl i.e. your current process does not contain mounts
1140198521 M * oliwel it contains mounts . at least I can access and see the mounted block devices
1140198548 M * Bertl well, obviously not inside xid=42
1140198560 M * Bertl chcontext --xid 42 --secure -- /bin/bash
1140198563 M * oliwel What is xid 42 ?
1140198567 M * Bertl (then you can look around)
1140198578 M * Bertl just an arbitrary context
1140198594 M * oliwel hm ok...
1140198599 M * oliwel but the disks are there...
1140198709 J * fwl ~fwl@83.215.237.1
1140198896 M * oliwel Bertl: 
1140198914 M * oliwel Bertl: self/mounts is there - just the symlink is missing....
1140198967 M * Bertl showattr /proc/mounts
1140198984 M * oliwel inside or outside the guest ?
1140198987 M * Bertl host
1140199001 M * oliwel AwH-ui- /proc/mounts
1140199008 M * Hollow that is a bug in util-vserver... it does not unhide /proc/mounts by default, and on gentoo we hide it by default..
1140199012 M * Hollow will fix it asap
1140199022 M * oliwel ARGH
1140199027 M * Bertl :)
1140199057 M * Bertl okay, so let's focus on the remaining issues
1140199060 M * oliwel Bertl: hjow can I unhide ?
1140199076 M * oliwel perhaps the whole unmounting is caused by the missing proc/monuts...
1140199097 M * Bertl setattr --~hide /proc/mounts
1140199134 M * Bertl oliwel: no, if your guest is gone, and the namespace was removed, there can be no reference to your device, except from the host
1140199173 M * oliwel hm...
1140199182 M * oliwel I reboot and will retry...
1140199322 Q * fwl Quit: This computer has gone to sleep
1140199339 J * fwl ~fwl@83.215.237.1
1140199383 M * oliwel Bertl: ok - /proc/mounts is there, but the device isn't free#d
1140199542 J * liquid3649_ ~Viper0482@p54976330.dip.t-dialin.net
1140199592 P * fwl 
1140199601 J * fwl ~fwl@83.215.237.1
1140199647 M * Bertl oliwel: okay, could you try with the holders before you start the guest, when the guest is started and afterwards?
1140199657 M * Bertl (all after a clean/new reboot)
1140199880 M * oliwel Bertl: already did this - free / busy / busy
1140199951 M * Bertl how busy?
1140199967 M * Bertl and I'm interested in the hex values
1140199977 Q * Viper0482 Ping timeout: 480 seconds
1140200103 M * oliwel ok mom
1140200125 M * oliwel bdev[f74d5100]: holder=00000000[0]
1140200127 M * oliwel bdev[f74d5100]: holder=c03b9b00[2]
1140200128 M * oliwel bdev[f74d5100]: holder=c03b9b00[2]
1140200145 M * Bertl are you sure the context was disposed at all?
1140200160 M * Bertl i.e. /proc/virtual/* doesn't show anything?
1140200206 M * oliwel mom
1140200252 M * oliwel there is a info and status :)
1140200316 N * ebiederm_zZ ebiederm
1140200326 M * oliwel mom on the phone
1140200337 M * Bertl morning ebiederm!
1140200347 M * ebiederm Morning Bertl.
1140200513 M * ebiederm Ugh a process without a a /proc/<pid>/mounts.
1140200601 M * ebiederm Now that I am caught up on my sleep I think I am going to see what I can do about getting my code ready, and start merging...
1140200787 M * Bertl ebiederm: we found the proc/<pid>/mounts :)
1140200802 M * ebiederm Good.
1140200808 Q * liquid3649_ Remote host closed the connection
1140200821 J * Viper0482 ~Viper0482@p54976330.dip.t-dialin.net
1140200985 M * oliwel Bertl: 
1140201002 M * oliwel sorry was urgently on the phone...for what sould I look in /proc/virtual
1140201051 M * Bertl well, first, for remaining contexts, but the status/info would be interesting too
1140201111 M * oliwel Bertl: there is only one file "info" and one "status" , no more
1140201119 M * oliwel cat info
1140201120 M * oliwel VCIVersion:     0002:0001
1140201122 M * oliwel VCISyscall:     273
1140201123 M * oliwel VCIKernel:      03000116
1140201129 M * oliwel cat status
1140201130 M * oliwel #CTotal:        0
1140201132 M * oliwel #CActive:       0
1140201143 M * Bertl okay, so no contexts are active
1140201188 M * oliwel Any way to track down the info from "holder" ?
1140201214 M * Bertl you mean, who is holding a reference?
1140201242 M * oliwel yes
1140201257 M * Bertl that's hard to tell, we could add some kernel logging for that
1140201266 M * oliwel sounds ugly
1140201271 M * oliwel but if it helps
1140201280 M * Bertl well, probably we have not many other chances than this :)
1140201293 M * oliwel Bertl: would it help to upgrade to 2.6.16 ?
1140201330 M * oliwel or use 2.1.1 instead 2.0.1 ?
1140201347 M * oliwel sorry - its 2.1.1 currently...
1140201413 M * Bertl while I prepare and test a log patch, you can test with 2.6.16 and/or 2.0.2
1140201459 M * oliwel My original kernel was 2.6.15 with 2.0.1 - so I guess that this is not the thing...
1140201474 M * Bertl okay, then let's try a 2.6.16 one
1140201490 M * oliwel vanilla with vs patches ? Or direclty gentoo ?
1140201502 M * Bertl vanilla + patches
1140201511 M * Bertl (until we know what we are looking for)
1140201519 M * oliwel ok - will take arrpund 5 mins
1140201538 M * oliwel 2.1 or 2.0 ? Any prefs ?
1140201590 M * Bertl let's keep 2.1.1 
1140201594 M * oliwel ok
1140201604 M * oliwel this is latest ? http://vserver.13thfloor.at/Experimental/patch-2.6.16-rc2-vs2.1.1-rc3.diff
1140201633 J * shedi ~siggi@inferno.lhi.is
1140201635 M * Bertl yup
1140201646 M * oliwel ehhm . I have rejects !
1140201684 M * oliwel Do you have a patchset againt rc3 already ?
1140201694 M * Bertl no, not yet
1140201702 M * oliwel ok so I will fetch rc2
1140201741 M * Bertl yup, won't hurt
1140201778 M * Bertl could you explain the block layers you ahve to me?
1140201808 M * oliwel Bertl: ??
1140201814 M * oliwel You mean my devices ?
1140201880 M * Bertl yup, dm/md/raw
1140201884 M * oliwel ok
1140201923 M * oliwel two P-ATA disks - each with 15 partitions, they form 15 md raid1 devices (md1 to md15)
1140201959 M * oliwel md1 to md4 are system disks - used as md block device
1140201977 M * oliwel md4 contains the basic vserver systems - it is read-only
1140202019 M * oliwel md5 to 15 are each part of a drbd device
1140202038 M * oliwel they hav a corresponding part on the second machine (100% clone)
1140202052 M * oliwel the drbd device  contain the data/configuration files for individual guests - one per guest
1140202056 M * oliwel thats it
1140202060 J * dothebart ~willi@xdsl-213-196-221-23.netcologne.de
1140202215 M * oliwel ok - I am ready...
1140202254 M * Bertl but the md5-15 are not used atm?
1140202300 M * oliwel for our testing - no 
1140202312 M * oliwel I used one of them as block device for the mount tests
1140202339 M * oliwel currently only 4 md devices are active (boot, root, var, vservers)
1140202372 M * oliwel drbd is not in the kernel at all and drbd helper tools are not started
1140202406 P * fwl Leaving
1140202517 Q * tudenbart Ping timeout: 480 seconds
1140204218 M * oliwel Bertl: vanilla 16rc2 - doesnt look better: http://pastebin.com/559985
1140204280 J * robig ~robig@83.221.253.158
1140204284 M * robig re
1140204305 M * robig hi, can somebody help me with my vservers exim configuration?
1140204319 M * Bertl robig: yup, probably the exim folks, no?
1140204333 M * robig where can i find them? :)
1140204362 M * robig can/shall i use one exim to send mails for all other vservers?
1140204449 M * Bertl http://www.exim.org/
1140204466 M * Bertl (mailing list and archive)
1140204505 M * oliwel Bertl: off for 10 mins...
1140204512 M * Bertl k
1140205222 Q * shuri Ping timeout: 480 seconds
1140205230 Q * Viper0482 Ping timeout: 480 seconds
1140205281 J * shuri ~shuri@64.235.209.226
1140205370 Q * rs Quit: rs
1140205657 M * Bertl oliwel: http://vserver.13thfloor.at/Experimental/delta-claim-test01.diff
1140206208 M * nayco Bertl: As I'm going to have some time again to experiment with vserver, in a couple of weeks, I'd like to know if you have some thing that need testing with the kernel ? And/or the utils ?
1140206227 M * nayco +s+s
1140206257 M * Bertl sure, the new scheduler and the CoW link breaking needs some testing
1140206336 J * dearaujo ~dan@pixpat.austin.ibm.com
1140206439 M * Bertl welcome dearaujo!
1140206451 M * nayco Mmmmm, more on that ? The CoW link breaking is about unification, I think, and what about the scheduler ?
1140206454 M * dearaujo good afternoon/evening
1140206481 J * djudko ~davej@129.33.1.37
1140206488 M * Bertl welcome djudko!
1140206525 M * Bertl nayco: well, nothing special to test, maybe isolation in general?
1140206532 J * xye ~xye@c-24-63-24-247.hsd1.ma.comcast.net
1140206541 M * Bertl welcome xye!
1140206616 M * nayco Ok. It'll be the occasion for me to try unification, as I never used it before... Oh, and what about "inside" disk quotas ?
1140206656 M * xye hi
1140206749 M * xye im trying to track down a problem that i found someone was having on this channel back in 2004
1140206761 M * xye basically getting pam_limits errors in the guest if it is started from the host init
1140206779 M * xye he seemed to track it down to a PRIO_PROCESS of 2 during the init but im not quite sure im seeing the same thing
1140206791 M * xye it was rs talkking about it back on 2004-7-22
1140206804 M * xye anyone remember this remember? Bertl you were helping him debug it
1140206822 M * Bertl nayco: not implemented on shared disks (for 2.6)
1140206835 M * oliwel Bertl: compiked and booting...
1140206849 M * Bertl xye: well, what log entries, and/or strace do you see?
1140206862 M * xye w
1140206864 M * xye ack
1140206885 M * xye pam_limits[6434]: setrlimit limit #6 to soft=-1, hard=-1 failed: Operation not 0
1140206899 M * xye truncated by my terminal, but it is operation not permitted
1140206935 M * xye http://64.233.179.104/search?q=cache:Fs4pb7bdxsYJ:www.sculpturedlife.com/vserver/irc-logs/index.php
1140206939 M * Bertl well, your pam is trying to set the limits to unlimited
1140206952 M * xye yea
1140206959 M * Bertl at the same time, it seems that your guest/startup is limited
1140206963 M * nayco Bertl: "On shared disks" ? Er... You mean "not possible to set quotas for a single partition, only for the whole vserver and/or disk" ?
1140206977 M * xye i can disable pam_limits but it is strange because this only happens if i start the guest from init in the host, if i do a vserver guest restart the problem goes away
1140206996 M * Bertl nayco: means you cannot have quota (but you can have disk limits) per guest on a single (shared) partition
1140207017 M * Bertl xye: probably your 'init' applies certain limits?
1140207040 M * xye Bertl, im just using plain with nothing setting any limits
1140207061 M * Bertl the host init, the one which starts the guest
1140207061 M * xye im using the vserver-default script to start them up on boot
1140207109 M * Bertl add a line to the startup script which outputs the currently set limits
1140207120 Q * nayco Quit: Bonne nuit !
1140207131 M * Bertl xye: 'ulimit -a'
1140207150 M * xye Bertl, startup script being the one that calls vserver-wrapper? (vservres-default)
1140207165 M * Bertl yup, or the vserver-default script
1140207177 J * Smutje_ ~Smutje@xdsl-87-78-2-179.netcologne.de
1140207225 M * oliwel Bertl: The system is online fresh boted, log-patch installed - what know...
1140207263 J * Viper0482 ~Viper0482@p54976330.dip.t-dialin.net
1140207284 Q * Smutje Ping timeout: 480 seconds
1140207284 N * Smutje_ Smutje
1140207436 J * nayco ~nayco@lns-bzn-29-82-64-225-98.adsl.proxad.net
1140207470 M * xye Bertl, good call, the ulimit differs during the init process
1140207471 M * Bertl oliwel: just do what you always do, we then have a look at the kernel log
1140207484 M * oliwel ok
1140207502 M * Bertl xye: okay, so if you can fix that, it should be fine
1140207503 M * nayco Bertl: Well, I remember what we did months ago : Quota worked, but with the help of vroot ...
1140207557 M * Bertl nayco: on separate partitions (for each guest) it works quite fine
1140207571 M * oliwel Bertl: same as ever - kernel log is empty :(
1140207582 M * Bertl dmesg?
1140207593 M * Bertl you sure you applied the patch?
1140207607 M * nayco Bertl: Oh, you mean with or without vroot ?
1140207636 M * oliwel Bertl: yepp - will crosscheck
1140207639 M * Bertl nayco: vroot is required for secure quota operation
1140207689 M * oliwel Bertl: can confirm - patch is applied
1140207698 M * nayco Ok, so, each quota controled partition needs a manual setup with vroot before using "quota" tools, right ?
1140207716 M * xye Bertl, is there a clean way to do this without modifying the util-vserver scripts? 
1140207725 M * Bertl nayco: well, it need vroot to make some quota tools work
1140207742 M * Bertl xye: no idea, it's your init doing that, no?
1140207757 M * xye Bertl, its whatever debian is doing combined with util-vserver packages
1140207763 M * nayco Bertl: Ok, thanks ;-)
1140207768 M * oliwel Bertl: should I do a "make clean" before re-compiling ? 
1140207778 M * Bertl xye: yeah, so well, ask the debian maintainer of init :)
1140207801 M * xye :)
1140207813 M * Bertl oliwel: no, it should rebuild some parts of the kernel
1140207821 M * nayco Bertl: Oh, and what about "vserver-utils-1.0.1" ?
1140207822 Q * lilalinux Remote host closed the connection
1140207843 M * Bertl nayco: IIRC they are at 1.0.3 or so
1140207851 M * oliwel Bertl: yes it did so, but dmesg just shows the three lines from holder
1140207868 M * Bertl oliwel: the patch was tested, it writes the following for every bd claim/release
1140207874 M * nayco Bertl: Er, I mean, do the now replace util-vserver-0.30.x ?
1140207880 M * Bertl oliwel: bd_claim: 8138d040 80481820 @fs/block_dev.c:896
1140207886 M * Bertl oliwel: bd_release: 8138d040 @fs/block_dev.c:918
1140207942 M * Bertl oliwel: unless you disabled kernel logging or you are using the wrong kernel you should get one for every mount done
1140207962 M * oliwel where can i enable/disable kernel logging ?
1140207981 M * oliwel teh kernel is the right one
1140207992 M * Bertl well, on the kernel command line (e.g. quiet) or via sysctl
1140208052 M * oliwel hmm - cant remember that I did so...
1140208062 M * oliwel and the holder loclks are shown too
1140208085 M * Bertl that's why I suspect that your kernel isn't the right one
1140208107 M * oliwel regarding the build date it is... uname -a
1140208109 M * oliwel Linux automatix 2.6.16-rc2-vs2.1.1-rc3 #2 Fri Feb 17 21:10:46 CET 2006 
1140208146 M * Bertl nm vmlinux | grep bd_claim
1140208190 M * oliwel nm: vmlinuz-2.6.16rc2-vs: File format not recognized
1140208206 M * Bertl the vmlinux in your build tree :)
1140208211 M * oliwel ok
1140208214 M * oliwel sorry
1140208242 M * oliwel 3 symbols found
1140208251 M * Bertl fine, which ones? :)
1140208258 M * oliwel c014c642 T __bd_claim
1140208260 M * oliwel c039a3e3 r __kstrtab___bd_claim
1140208262 M * oliwel c0393944 r __ksymtab___bd_claim
1140208278 M * Bertl that looks fine, so why does the printk() not log?
1140208289 M * oliwel Dont ask me...
1140208300 M * oliwel its a default vanilla kernel
1140208316 M * Bertl well, somehow I get the feeling your system is borked :)
1140208330 M * oliwel # CONFIG_DEBUG_KERNEL is not set ??
1140208340 M * Bertl doesnt matter
1140208351 M * oliwel Hmm it worked till last week . I just did a gentoo update...
1140208357 M * oliwel no more
1140208385 M * Bertl the holder stuff does nothing different (regarding printk) than the bd_claim
1140208406 M * Bertl now, assuming that your kernel was built properly and the bzImage updated
1140208429 M * Bertl and you booted from there, then this would mean that no block device is claibed
1140208433 M * Bertl *claimed
1140208453 M * Bertl this, OTOH, would result in the holder patch always printing zero
1140208543 M * oliwel hmmm
1140208553 M * oliwel And I can confirm that I have some blkdev....
1140208571 M * oliwel Should I try a make clean and rebuild ?
1140208586 M * Bertl well, you can try ...
1140208601 M * Bertl the kernel build system should handle that properly though ...
1140208613 M * oliwel Dont think taht it matters...but I dont have any other idea....you ?
1140208626 M * Bertl nope
1140208696 M * oliwel if that wont worlk I thik I will start the system from scratch
1140209254 M * oliwel Bertl: System is back
1140209262 M * oliwel Now I have the logs in the kernel
1140209293 M * Bertl good :)
1140209358 M * oliwel http://pastebin.com/560134
1140209370 M * oliwel Think I just missed the lines...
1140209377 M * oliwel expected more ;)
1140209539 M * Bertl okay, what happens if you unmount the newly mounted dir?
1140209540 M * oliwel Bertl: http://pastebin.com/560134
1140209558 M * oliwel ergh - what should I unmount ?
1140209572 M * Bertl the ext3 fs which was obviously mounted
1140209593 M * oliwel I cant unmount it...
1140209610 M * oliwel its not there anymore...
1140209614 M * Bertl well, what happens if you try?
1140209628 M * oliwel umount /dev/md14
1140209629 M * oliwel umount: /dev/md14: not mounted
1140209692 M * Bertl what does your fstab (for the guest) contain?
1140209704 J * apg88 ~apg88@adsl-065-007-211-201.sip.mia.bellsouth.net
1140209722 M * Bertl welcome apg88!
1140209772 M * oliwel http://pastebin.com/560140
1140209826 M * dearaujo hello, if I was going to run a single webserver on multiple instances of a guest, I would assume that assigning private IPs and using nat,dnat would be cumbersome (if not impossible)? opinions?
1140209828 Q * apg88 Quit: Leaving
1140209861 M * Bertl dearaujo: how would you run a single webserver on multiple instances of a guest?
1140209886 Q * Viper0482 Quit: bin raus, 
1140209894 M * dearaujo Bertl: 1 webserver inside each guest
1140209903 M * dearaujo Bertl: can that not be done?
1140209921 M * Bertl sure, just the 'wording' was confusing
1140209929 M * dearaujo oh, I'm sorry :)
1140209958 M * Bertl dearaujo: if you have public ips for each of them, then no S/NAT is required
1140209977 M * dearaujo yes but if I had 1 public IP
1140209978 M * Bertl dearaujo: if you 'share' one IP for all of them, you need some proxy which understands http
1140209998 M * Bertl i.e. some master httpd which distributes the requests
1140210027 M * dearaujo ah ok - so it must be done at a higher level than s/nat
1140210038 M * oliwel Bertl: http://pastebin.com/560140 - this was for you ;)
1140210040 M * Bertl oliwel: what happens if you remove that /data mount?
1140210056 M * dearaujo Bertl: thanks
1140210081 M * Bertl dearaujo: yep, because http first conencts, then sends the url
1140210107 M * Bertl dearaujo: be aware that ssl requires different IPs
1140210122 M * Bertl dearaujo: well, at least for per domain certificates
1140210128 M * oliwel Bertl: how ??
1140210138 M * oliwel You mean from the fstab ?
1140210144 M * Bertl oliwel: yup
1140210204 M * oliwel In this case the device is never mounted...
1140210212 M * oliwel and I think I wont get any trouble
1140210268 M * oliwel So it is with the other md devices that are unued - they stay as they are... - so I dont nuderstand your suggeston - the device is NOT mounted outside the guest
1140210519 J * bonbons ~bonbons@83.222.39.180
1140210549 M * Bertl oliwel: the log shows an ext3 mount, where does that come from, in your opinion?
1140210604 Q * djudko Quit: Leaving
1140210621 M * oliwel Bertl: When I start the guest, the /dev/md14 is mounted INSIDE the guest from the block device - this is the ext3 mount
1140210635 M * Bertl precisely
1140210656 M * oliwel this mount exists onyl in the guests-context, so it should be released when the context dies
1140210657 M * Bertl I suspect it is never unmounted (or at least it doesn't drop the claim)
1140210670 M * oliwel I agree....but why
1140210690 M * Bertl that is something we have to investigate, once we know that this _is_ the issue
1140210706 M * oliwel I will do my best to help you...
1140210722 M * Bertl k, so let's comment it out then :)
1140210724 M * oliwel It seems that everything is fine on the old system...this was a 2.6.13 with 2.0.1...
1140210737 M * oliwel I have some time left ;)
1140210744 M * Bertl I suspect a mainline issue
1140210752 M * oliwel ouch
1140210774 M * Bertl will update to 2.6.16-rc3 in the meantime
1140210833 J * Doener doener@i5387EB11.versanet.de
1140210835 M * oliwel the patches will not work...
1140210840 M * Doener evening!
1140210847 M * Bertl evening Doener!
1140210873 M * oliwel Bertl: anything I can do now ?
1140210888 P * robig 
1140210893 M * oliwel My schedule is hard - I should have the system stable till monday ;(
1140210924 M * Bertl oliwel: yep, did you try withtout the /data mount yet?
1140210940 M * oliwel What do you expect ?
1140210988 M * Bertl ah, hmm, now I udnerstan
1140210992 M * Bertl *understand
1140211003 M * Bertl your problem _is_ the md14 mount, right?
1140211021 M * oliwel Bertl: Yes
1140211052 M * oliwel Bertl: Tried: mount md13 outside the guest start the guest, stop the guest, free the mount - works
1140211125 M * oliwel so it looks for me as the mount done inside the guest is not released properly
1140211268 M * oliwel Bertl: When I mount the device like usual within guets's fstab and unmount it with "vnamespace....umount" before shuting down - it works !!!
1140211292 M * oliwel So - I asume that the utils are broken !!!
1140211433 M * Doener no, that'd probably be a kernel problem
1140211442 M * Doener when the namespace is gone, the mounts should be as well
1140211445 M * Hollow dunno if the umount function is borken in util-vserver, but shouldnt the mount be released when the namespace disappears?
1140211450 M * Hollow ahw ell
1140211473 M * oliwel Hollow: i am doing a downgrad to 209..lets see...
1140211514 M * bonbons oliwel: did you check that the context was correctly released (e.g. not persistent)
1140211543 M * Hollow util-vserver has no support for persistant
1140211556 M * Hollow which should be persistent btw..
1140211623 M * oliwel Hollow: why does my guest fail with older utils ?
1140211628 M * oliwel chbind: vc_set_ipv4root(): Invalid argument
1140211630 M * bonbons never looked so deeply at it, moreoverthe util-vserver I have running is not the last version (so persistence might have been introduced since then)
1140211639 P * dearaujo 
1140211718 M * Bertl oliwel: dynamic contexts disabled
1140211799 M * oliwel In the kernel . yes
1140211804 M * oliwel is this the problem ?
1140211851 M * oliwel Bertl: seems I cant start the vserver with the downgraded Tools :(
1140211869 M * Doener oliwel: which kernel/vserver version are you running?
1140211925 M * oliwel kernel: everything between 2.6.15.4 and 2.6.16rc2 gentoo and vanilla in the last few hours
1140211942 M * oliwel vserver .210 - downgraded to 209 to test, but wont work
1140212000 M * oliwel Bertl: should I enable dynamic context ?
1140212025 M * Doener vserver == kernel patch, not tools ;) 
1140212051 M * oliwel Doener: latest 2.1.1 against 15.4 and 16rc2
1140212067 M * oliwel but also tried 2.0.1.4
1140212657 M * Bertl oliwel: could you try the following: vnamespace --new -- mount /dev/md14 /mnt
1140212717 M * Bertl oliwel: ah, and check the logs :)
1140212735 Q * nayco Quit: Bonne nuit !
1140212736 M * oliwel Bertl: mom - rebooting...for what ist "dynmaic context"?? I enabled it now in my kernel 
1140212771 M * Bertl yup, seems like ext3 has a dangling ref
1140212802 M * Bertl (ext2 unmounts properly, ext3 keeps the ref)
1140212842 M * oliwel you mean that it depends on the fs.... I had reiser before...
1140212850 M * Bertl yup, ext2 seems fine
1140212883 M * oliwel ok - I mounted..what now ?
1140212897 M * Bertl check if the device is claimed
1140212941 M * oliwel seems so 
1140212943 M * oliwel bd_claim: f7d97ac0 c039e680 @fs/block_dev.c:896
1140212945 M * oliwel kjournald starting.  Commit interval 5 seconds
1140212946 M * oliwel EXT3 FS on md14, internal journal
1140212948 M * oliwel EXT3-fs: recovery complete.
1140212949 M * oliwel EXT3-fs: mounted filesystem with ordered data mode.
1140212967 M * Bertl the important detail is the missing bd_release :)
1140213015 M * oliwel how can i determine the context id choosen ?
1140213105 M * Bertl no context id, this is purely mainline
1140213135 M * Bertl would be interesting to try this with vanilla + bdclaim patch (but without vserver)
1140213136 M * oliwel ah ok..
1140213162 M * oliwel Bertl: IIUC - its an ext3 issue ?
1140213228 M * Bertl looks like, might be cause be the vserver patch though
1140213251 M * Doener IIUC?
1140213300 M * oliwel IIUC - if I undestand correctly
1140213313 M * oliwel affirmative...using md14 with reiser works fine
1140213334 M * oliwel F++ck
1140213398 M * oliwel so the problem is vanilla - vserver - ext3
1140213419 M * Bertl we don't know yet, maybe it's purely vanilla
1140213471 M * oliwel ok - unmounting on vanilla is done not so often...
1140213486 M * oliwel at least it means that I have to use != ext3
1140213490 M * Bertl you can try the same command sequence with vanilla
1140213494 M * oliwel Bertl: what are you using ?
1140213560 M * oliwel will vnamespace work with vanilla
1140213604 M * Bertl I'd say so
1140213637 M * oliwel Hollow: didnt you check in 2.6.16 ?
1140213652 M * phreak`` oliwel: nope I didn't ..
1140213654 M * oliwel hmm.. will you or should I ive it a try ?
1140213662 M * oliwel phreak``: ok...
1140213711 M * phreak`` oliwel: If you wanna give it a try, just say so and I'll prep a tarball for that
1140213745 M * oliwel phreak``: not so important...will use 15.4 so long
1140213760 M * phreak`` k :)
1140213979 M * oliwel phreak``: what alternative to ext3 would you suggest
1140213981 M * oliwel xfs ?
1140214010 M * Bertl oliwel: no, we are going to fix the ext3 issue
1140214037 M * oliwel ok - you wil have my suport 
1140214425 Q * oliwel Quit: Chatzilla 0.9.69.1 [Firefox 1.5/2005111116]
1140214442 Q * xye Remote host closed the connection
1140215755 J * oliwel ~chatzilla@host-62-245-151-178.customer.m-online.net
1140215779 J * fh fiona@cheeto.net
1140215977 Q * cehteh Ping timeout: 480 seconds
1140216025 M * oliwel Bertl: so - back - if I can assist you hunting the ext3 bug, tell me how..
1140216050 M * Bertl currently compiling a vanilla kernel + claim patch
1140216077 M * oliwel ok
1140217165 Q * monrad Ping timeout: 480 seconds
1140217420 Q * bonbons Quit: Leaving
1140218970 M * Bertl seems like it _is_ a mainline issue ...
1140219461 P * meandtheshell 
1140219508 M * Bertl compiling here takes agaes :)
1140219514 M * Bertl *ages even
1140219617 Q * shuri Ping timeout: 480 seconds
1140219624 M * derjohn Bertl, latest devel for 2.6.15 still .11 ?
1140219638 M * Bertl hmm?
1140219664 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.15.4-vs2.1.1-rc6.diff
1140219672 M * derjohn i mean the lastest devel that compiles on 2.6.15 ... 2.1.0.11 ... 
1140219677 M * derjohn uercks?
1140219691 M * derjohn (German word of being surprised)
1140219697 M * Bertl lol
1140219725 M * derjohn 2.1.1-rc6 ? sill a long way to "2.1.1" ?
1140219743 M * Bertl well, as it seems we have to fix a few mainline issues :)
1140219756 M * derjohn with 2.6.16 only?
1140219767 M * Bertl nope, at least 2.6.15 and 2.6.16
1140219775 M * derjohn or do you refer to the mainline integration?
1140219785 M * Bertl no, brokeness with ext3
1140219797 M * derjohn hm, maybe I should stay an 2.6.14 in production?
1140219809 M * Bertl oliwel: did you test with 2.6.14 btw?
1140219816 M * derjohn risk of data loss????
1140219828 M * Bertl always :)
1140219836 M * derjohn any particular issue?
1140219844 M * Bertl no, it's a reference which should not be held
1140219854 M * Doener Linus must have noticed that i just downloaded rc3... ;)
1140219862 M * derjohn incontinentia referencia ?
1140219864 M * derjohn ;/
1140219869 M * Bertl ah, rc4 is out? *G*
1140219875 M * Doener exactly ;)
1140219902 M * Doener ext3 issue is not mentioned in short log
1140219925 M * derjohn well folks, is there any of {eric|sam|dhans..|etc} mainline effort already in rc4 ?
1140219962 M * derjohn this is goinf to be more confusing then before ... with all that .. hm .. lets call it branches or even slices ...
1140220012 M * derjohn Bertl, I always told you to take reiser ;) [but lately changed to XFS due to fear on reiser4]
1140220063 M * Bertl Doener: if you want to test, it's quite easy
1140220075 M * Bertl vnamespace --new -- mount /dev/somedev /mnt
1140220097 M * Bertl then check (e.g. with holders tool) for exclusive open of somedev
1140220167 J * shuri ~shuri@64.235.209.226
1140220271 M * Doener not today, no sleep last night, i.e. going to bed now :) good night everyone!
1140220278 M * Bertl night then!
1140220288 M * derjohn n8 Doener !
1140220293 Q * Doener Quit: Leaving
1140220351 M * oliwel Bertl: I tried 2.6.16 during our test, but not in production - now I am back to 15.4 because its the latest one thats in portage
1140220406 M * Bertl well, as far as I can tell 2.6.15 has this issue
1140220410 M * oliwel Bertl: what timeline do yopu think for the fix - today, tomorrow, a week ?
1140220423 M * Bertl oliwel: so the question is, did it work for you with ext3 at all?
1140220450 M * oliwel Bertl: having this knowledge it perfectly explains why I had this trouble in the past
1140220463 J * monrad ~mikkel@213083190131.sonofon.dk
1140220474 M * oliwel IIRC I had one hanging mountpoint in the past that was the only ext3 between lots of reisers
1140220499 M * oliwel so I assume that my 2.6.13 running on the old production machine already has this bug
1140220523 M * Bertl hmm, could you test there with the vnamespace and holders tool?
1140220547 M * oliwel Bertl: no - as the current machine is the second node of this one I cant :)
1140220561 M * oliwel I can install 2.6.13 on the machine we are xurrently testing on..
1140220573 M * oliwel so we will see
1140220578 M * Bertl okay, let's try that
1140220657 M * derjohn well, I decided to wait with buling my next kernel package. with 2.6.14/2.1.0.5 if did not have any issues yet.