1177200520 J * Loki|muh_ loki@satanix.de 1177200521 Q * Loki|muh Read error: Connection reset by peer 1177200530 N * Loki|muh_ Loki|muh 1177201435 Q * meandtheshell Quit: Leaving. 1177203394 J * FCOJ ~mordur@dsl-201-4.hive.is 1177203590 Q * FCOJ 1177203610 J * FCOJ ~mordur@dsl-201-4.hive.is 1177204062 J * tzafrir_laptop ~tzafrir@bzq-179-75-202.static.bezeqint.net 1177204086 Q * FCOJ Quit: Leaving 1177204152 M * tzafrir_laptop any rtfm on how to install centos5 on my vserver host? 1177205046 J * DoberMann_ ~james@AToulouse-156-1-2-168.w81-250.abo.wanadoo.fr 1177205074 M * Bertl_vV tzafrir_laptop: on the host? via the install CD I'd say ... 1177205105 M * tzafrir_laptop I was hoping to avoid the download of the 6 CDs 1177205128 M * Bertl_vV host is what we call the physical machine ... 1177205143 M * Bertl_vV guest is what we call the isolated units 1177205146 M * tzafrir_laptop right. I have a running host with several guests 1177205153 Q * DoberMann[ZZZzzz] Ping timeout: 480 seconds 1177205206 M * tzafrir_laptop call them what you want. I won't use such terms until you come up with ones as short as "host" and "guest :-p 1177205249 M * Bertl_vV hehe 1177205263 M * Bertl_vV okay, so you still want to reinstall the host? or a guest? 1177206030 M * tzafrir_laptop add a new guest 1177206099 M * Bertl_vV something like this should work 1177206102 M * Bertl_vV vserver min-centos build -m yum --context 42 --hostname my.domain.com --interface domain=eth0:192.168.0.136/24 --initstyle sysv \ -- -d centos 1177206110 M * Bertl_vV hmm, ignore the \ 1177206144 M * Bertl_vV vserver min-centos build -m yum --context 42 --hostname my.domain.com --interface eth0:192.168.0.136/24 --initstyle sysv -- -d centos 1177206174 M * Bertl_vV you might need to change centos to centos5 and/or configure that distro properly 1177206186 M * Bertl_vV (depending on the tool version) 1177206188 M * tzafrir_laptop are you sure it works for centos5? I have "centos4" type in my configurations. I don't recall any type "centos" 1177206231 M * tzafrir_laptop 0.30.212-1 (from Debian Etch) 1177206235 M * Bertl_vV I think latest tools (cvs) should have a centos5 too 1177206257 M * Bertl_vV but you might create one yourself 1177206272 M * Bertl_vV (by adapting the centos4 one) 1177206281 M * tzafrir_laptop noone put a newer version in Debian Unstable yet... 1177206322 M * Bertl_vV yeah, debian always is somewhat behind ... 1177206401 M * Bertl_vV okay, off for bed now .. good luck ... 1177206615 Q * DreamerC_ Quit: leaving 1177206631 J * DreamerC ~dreamerc@125-225-100-122.dynamic.hinet.net 1177206669 M * tzafrir_laptop well, 0.30.212 is the recent version according to linux-vserver.org . Anyway, thatnks 1177206868 M * tzafrir_laptop right. centos5 was added to the svn 12 hours ago 1177207946 Q * bzed Quit: Leaving 1177209308 Q * ensc Ping timeout: 480 seconds 1177211333 J * ktwilight_ ~ktwilight@31.122-66-87.adsl-dyn.isp.belgacom.be 1177211743 Q * ktwilight Ping timeout: 480 seconds 1177212198 J * ensc ~irc-ensc@p54B4E9E3.dip.t-dialin.net 1177213568 Q * infowolfe Read error: Connection reset by peer 1177214542 J * infowolfe ~infowolfe@c-67-164-195-129.hsd1.ut.comcast.net 1177214583 Q * infowolfe 1177214778 J * infowolfe ~infowolfe@c-67-164-195-129.hsd1.ut.comcast.net 1177215367 M * mattzerah_ hi guys - anyone here experiened with xdmcp - especially gdm 1177215404 M * mattzerah_ i need to restrict the listening of my host xdmcp server to an address (instead of all addresses) so it doesn't interfere with my guest xdmcp 1177215419 M * mattzerah_ any hints would be very appreciated :) 1177216529 J * FireEgl ~FireEgl@adsl-147-76-85.bhm.bellsouth.net 1177217630 Q * tzafrir_laptop Ping timeout: 480 seconds 1177218074 J * infowolfe_ ~infowolfe@c-67-164-195-129.hsd1.ut.comcast.net 1177218082 Q * infowolfe_ 1177218242 J * infowolfe_ ~infowolfe@c-67-164-195-129.hsd1.ut.comcast.net 1177218309 Q * infowolfe Killed (NickServ (GHOST command used by infowolfe_)) 1177218314 Q * infowolfe_ 1177218326 J * infowolfe ~infowolfe@c-67-164-195-129.hsd1.ut.comcast.net 1177218488 Q * softi42 Ping timeout: 480 seconds 1177218555 M * Bertl_vV rmrmrf 1177218794 M * oliwel Bertl_vV: good morning ;) 1177219060 J * softi42 ~softi@p549D5F10.dip.t-dialin.net 1177222134 Q * infowolfe Quit: Leaving 1177223195 Q * Aiken Remote host closed the connection 1177226639 J * infowolfe ~infowolfe@c-67-164-195-129.hsd1.ut.comcast.net 1177230060 Q * FireEgl Ping timeout: 480 seconds 1177230335 J * meandtheshel1 ~markus@85-124-206-5.dynamic.xdsl-line.inode.at 1177230884 N * DoberMann_ DoberMann 1177231186 J * boci^ ~boci@pool-1834.adsl.interware.hu 1177233431 J * bonbons ~bonbons@83.222.38.145 1177233601 J * bragon_ ~bragon@ACaen-152-1-7-131.w86-220.abo.wanadoo.fr 1177233998 Q * bragon Ping timeout: 480 seconds 1177234008 J * bzed ~bzed@dslb-084-059-125-165.pools.arcor-ip.net 1177234978 J * pyquila gerbens@82.94.222.35 1177235363 M * oliwel knockknock ? 1177235373 M * daniel_hozac who's there? 1177235390 M * oliwel Hi daniel_hozac 1177235397 M * oliwel nice to see you ;) 1177235410 M * daniel_hozac hello 1177235434 M * oliwel it seems like a lot of stuff changed with my upgrade - almost everything has issues :( 1177235479 M * oliwel but it seems as it is not the vserver stuff but some filesystem utils from upstream have changed their behaviour 1177235489 M * oliwel but perhaps you have an idea 1177235502 M * oliwel I have a severe problem reagrding mounts 1177235521 M * daniel_hozac oh? 1177235545 M * oliwel i mount a virtual /dev node (tmpfs) and a physical drive (drbd disk) within the guests fstab supported by a bunch of scripts in the post/pre start dirs 1177235553 M * oliwel and nothing does like before 1177235566 M * oliwel it seems that mounts are not correctly cleaned up 1177235570 M * oliwel so - first 1177235598 M * oliwel I have a prepre-start script that create a tmpfs on /dev and populates it with necessary device nodes 1177235606 M * oliwel this works like beofre 1177235611 M * oliwel before 1177235650 M * oliwel the old setup cleaned up the node during shutdown 1177235661 M * oliwel the new setup behaves very strange 1177235689 M * oliwel the node appears in the root-context 1177235696 M * oliwel I can see it when I run "mount" 1177235717 M * daniel_hozac how do you mount the /dev inside the guest? from the prepre-start script? 1177235726 M * oliwel if I try to remove it, the system says the mount is not there and cleans the entrie from the mounttab 1177235730 M * oliwel yes 1177235743 Q * sladen Ping timeout: 480 seconds 1177235744 M * oliwel but I also tried oving it to fstab 1177235749 M * oliwel same result 1177235820 M * daniel_hozac you'd want to make your script pre-start if you put the mount in the guest's fstab. 1177235834 M * oliwel also tried this 1177235836 M * oliwel same result 1177235854 M * oliwel the setup of the node is ok 1177235872 M * oliwel but I am confused that it shows up in the root context 1177235882 M * daniel_hozac remove them then. 1177235884 M * oliwel it shouldnt be there 1177235891 M * oliwel Remove what ? 1177235909 M * daniel_hozac they would've been created when you started the guest with the mount in the fstab, and the script as prepre-start. 1177235915 M * daniel_hozac the device nodes. 1177235963 M * oliwel the setup worked on the old system without any problems.... 1177235978 M * oliwel and the nodes should not show up in the root host 1177236004 M * oliwel afaik all scripts are ran in the namespace of the guest 1177236007 M * daniel_hozac they only show up on the host because you experimented .) 1177236012 M * oliwel no 1177236019 M * oliwel I removed them and did a clean startup 1177236029 J * sladen paul@starsky.19inch.net 1177236040 M * daniel_hozac so you have the script as pre-start and the mount in the guest's fstab now? 1177236053 M * oliwel yes 1177236057 M * oliwel will paste 1177236059 M * oliwel mom 1177236075 M * oliwel in guets fstab 1177236078 M * oliwel > none /dev tmpfs size=100k 1177236092 M * oliwel in pre-start.d 1177236094 M * oliwel cp -pr /vservers/$2/_dev.tpl/* /vservers/$2/dev/ 1177236263 M * daniel_hozac you need to add ,dev 0 0 to that fstab line. 1177236378 M * oliwel 100k,dev 0 0 ? 1177236381 M * daniel_hozac yep. 1177236390 M * daniel_hozac is there a particular reason you have a tmpfs mount for /dev? 1177236397 M * oliwel Yep 1177236402 M * daniel_hozac i mean, since you're already copying the devices from the guest's root.. 1177236407 M * oliwel I share the guest rootfs 1177236422 M * daniel_hozac so? 1177236431 M * oliwel and otherwise the syslog socket gets "shared" between the guests 1177236447 M * oliwel the root is read-only 1177236475 M * oliwel the config stuff is mapped into the guests by a lot of symlinks and overlays from a per-guest config partition 1177236495 M * oliwel but thats all nothing new - I run this setup since 18 month now 1177236503 M * oliwel and it ran fine 1177236520 M * oliwel I just did a major update of the complete host system 1177236526 M * oliwel and now I have the probs 1177236589 M * oliwel shit - race condition now.... 1177236605 M * oliwel I mount a clean /dev and then I cant mount /dev/pts in the fstab..... 1177236664 M * daniel_hozac you have /dev before /dev/pts in the fstab, right? 1177236670 M * oliwel yep 1177236727 M * oliwel hm - if I swap it this means I use the same pts for all guests - unsure if this brings a problem 1177236855 M * oliwel does nto work . the later //dev overlays the pts so I cant acces it later 1177236934 M * oliwel any ideas ? 1177237136 M * oliwel hmm - interessting: I moved the creation of the dev to prepre-start again - if I do a "mount" I can see the node in the root context but not in the guest 1177237152 M * oliwel it seems that this is simial to my problem with mount I had with the nagos script 1177237169 M * oliwel the behaviour changed what mount displays regarding contexts/namespaces 1177237266 M * oliwel daniel_hozac: still there ? 1177237291 M * doener you shouldn't trust the "mount" output, it just reads /etc/mtab 1177237323 M * doener and your prepre-start script probably runs without a chroot, so mount writes its actions to the host's /etc/mtab 1177237339 M * oliwel doener: yes I know - I dont care about "mount" as a comamdn but the behaviour is strange all around issues regrding mounting 1177237375 M * oliwel hm, that might be true - but what confuses nme ist that this all worked before I updated my host system 1177237464 M * doener what about the inability to mount /dev/pts on a clean dev? what error messages did show up? 1177237519 M * oliwel its obvious - as the diretory /dev/pts does not exist I cannot mount anything there 1177237585 M * doener ah, totally clean tmpfs, right... 1177237655 M * doener hm, found a mount bug... 1177237677 M * oliwel ? 1177237702 M * doener unrelated... mount --move, does the same thing as --bind WRT the changes to /etc/mtab 1177237736 M * oliwel ah ok 1177237747 M * doener none on /root/ttt type tmpfs (rw) 1177237747 M * doener /root/ttt on /root/aaa type none (rw) 1177237748 M * oliwel so - you dont have any idea though what goes wrogn her ? 1177237764 M * doener after mounting tmpfs of /root/ttt and then moving it to /root/aaa 1177237781 M * tanjix doener: are you familiar with debian? 1177237806 M * doener tanjix: At least I'm running sid on both of my boxes ;) 1177237859 M * tanjix doener: hehe - maybe you can help me... i am running etch with the precompiled linux-image-2.6.18-4-vserver-686. but this only supports up to 4 gb ram - i need this kernel with up to 64 gb ram support 1177237875 M * tanjix doener: i tried lots aof things but i did not get it to work :( 1177237923 M * doener oliwel: if you mount it in your prepre-start script, does the mount only show up in the host's mtab, or also in /proc/mounts of the host? 1177237934 M * oliwel mom 1177237956 M * doener tanjix: I've never worked with debian kernels 1177237992 M * tanjix :( ok thx anyway :) 1177237995 M * oliwel doener: I can see it in hosts mtab and in the guest /proc..... 1177238017 M * doener so it's not in the host's /proc/mounts, right? 1177238041 M * oliwel nope 1177238065 M * doener do you need it to show up in the guest's /etc/mtab? 1177238081 M * oliwel hm - I guess not 1177238099 M * oliwel anyway - mtab will become a symlink to /proc mounts as the fs is readonly 1177238107 M * oliwel juts removed this for debuggin 1177238113 M * doener then you can just pass -n to mount to avoid the mtab entry 1177238125 M * oliwel in the prepare script ? 1177238165 M * doener guess so.. if that's the prepre-start thing that calls mount ;) 1177238176 M * oliwel yep it is - will try 1177238224 M * oliwel ok it did the trick 1177238240 M * oliwel now I will symlink mtab against /proc/mounts and all probs should be gone ;=) 1177238246 M * oliwel doener: time for another riddel 1177238249 M * oliwel riddle 1177238277 A * doener goes calling Batman to fight the evil Riddler 1177238300 A * oliwel calls Batgirl to assist doener 1177238315 M * oliwel ok 1177238321 M * doener hm, she might distract me, no? 1177238323 M * oliwel You know what drbd is ? 1177238340 M * doener some network blockdevice IIRC 1177238356 M * oliwel ok - the essentuial point 1177238368 M * oliwel you have the device on tow nodes but only one can be the master 1177238384 M * oliwel so - use this as data partiton inside the guest 1177238427 M * oliwel if the server is not running the device is in idle state, I must manually set it to "master" before I can mount it 1177238431 M * oliwel I do so in the prepre-start 1177238436 M * oliwel this worlds 1177238437 M * oliwel works 1177238443 M * oliwel I than mount it via fstab 1177238446 M * oliwel this works too 1177238465 M * oliwel when I shutdown the server the device gets unmounted by fstab again 1177238482 M * oliwel there is a postpost-stop script that should set the state back to idle 1177238495 M * oliwel and it fails with the msg "device is busy" 1177238519 M * oliwel if I rtun the idle command right after the stop script finished I can shutdown the disk 1177238546 M * oliwel I also tried a sleep in the shutdown script to allow some time to settle - but still the same 1177238641 M * oliwel Can you fight this evil riddle or will riddler take over gotham ? 1177238658 M * doener that question needs knowledge of how exactly util-vserver does its stuff, and as I'm still scared by the amount of bash involved, I lack that knowledge. Goodbye Gotham City! 1177238669 M * oliwel ct /proc/mounts shows the device is still mounted ...... 1177238677 M * doener guess gotham will get bash'ed ;) 1177238682 M * oliwel hm.... 1177238690 M * oliwel So I will call Superman 1177238701 M * doener and destroy Metropolis along the way? 1177238728 M * oliwel who cares about a city full of schmucks.... 1177238733 M * doener *lol* 1177238749 M * oliwel interesting: the device is still mounted according to /proc/mounts 1177238765 M * oliwel if I issue "unmount..." it tells me there is no deviec... 1177238813 M * doener did you use the path relative to the guest's or the host's root? 1177238824 M * oliwel i used teh device 1177238840 M * doener ah 1177238856 M * doener maybe /dev is gone before? 1177238875 M * oliwel anyway - according to the spec postpost is run after everythin is stopped - so the mounts should have been away 1177238882 M * oliwel hmmmm - thats a good point 1177238904 M * oliwel its a deadlock 1177238920 M * doener using umount $MOUNT_POINT should still work then 1177238932 M * oliwel mount point = directory node ? 1177238953 M * oliwel but anyway the question remains - why is the mount not cleaned up by init ? 1177238959 M * doener yeah, wherever that thing is mounted, i.e. the second column in /proc/mounts 1177238976 M * oliwel will try 1177239058 M * doener probably because util-vserver does the unmounting and doesn't know that is has to care about /dev being unmounted last 1177239089 M * oliwel any way the /dev inside the guest is not authorative for the block devices 1177239125 M * oliwel Hmm, umount /vserver/test/mountpoint does not show any erreor but the mount is still there (in proc/mounts) and I cant detach the device 1177239240 M * doener hm, could you http://paste.linux-vserver.org the contents of /proc/mounts of the host and guest while the guest is running? 1177239267 M * oliwel will do so... 1177239319 M * oliwel http://paste.linux-vserver.org/1544 1177239412 M * oliwel and 1177239414 M * oliwel http://paste.linux-vserver.org/1545 1177239511 M * doener hm, postpost-stop should have unmounted the directories already, at least according to the flower page... 1177239515 M * oliwel Strange that the mount is shown twice (line 32 / 38) 1177239519 M * oliwel yep 1177239526 M * doener that's the magic rbind 1177239536 M * oliwel as I said the setup was running fine for over 18 months now 1177239564 M * doener sounds like postpost-stop broke a bit 1177239579 M * oliwel I rebuild the host from scrathc and copied over the /etc/vserver tree 1177239594 M * oliwel I roled back the tools to the old version - still teh same 1177239605 M * oliwel so I guess its something related to upstream tools 1177239638 M * oliwel We had similar problems 2 y ago because a kernel handler was broken and did not celan up its mount-refs 1177239665 M * oliwel anyway - this is anoying as it breaks my whole auto-failover 1177239691 M * oliwel if the device is not unmounted I cant remount it on the other node 1177239765 M * oliwel otherwise - as I can shutdown the device immediately after the vserver script returns it seems that everything is cleaned up correctly 1177239785 M * oliwel it seems that the mount is not correctly cleaned from the mount list 1177239792 M * doener that's the auto cleanup after the namespace is destroyed 1177239837 M * doener if you "umount -n /vservers/test/data; umount -n /data;" does that help? 1177239838 M * oliwel hm - but the namespace should be already gone when postpost is runing 1177239886 M * oliwel trying 1177239894 M * oliwel no 1177239905 M * oliwel umount /data -> not found 1177239914 M * doener almost thought so... 1177239919 M * oliwel umount /vserver/test/data seems to execute without error 1177239937 M * doener how to /proc/mounts look after the successful umount? 1177239946 M * oliwel cat /prc/mounts still shows ONE ref to the device 1177239982 M * oliwel so - one is gone 1177239989 M * doener the /data one, right? 1177239998 M * oliwel yes 1177240026 M * oliwel so we cant kill this one.... 1177240028 M * oliwel crap 1177240035 M * doener hm, try: "umount -n /vservers/test/data; chroot . umount -n /data" 1177240041 M * doener no 1177240056 M * doener won't help, you don't reach the rbind'ed / that way 1177240091 M * oliwel but its not my job to clean up the mounts :( 1177240124 M * doener yeah, there's some evil bug in there 1177240139 M * oliwel do you think it might help if I mount the device from the scripts instead the fstab ? 1177240165 M * oliwel Guess we should wait for bertl - he fixed the last time the thing in minutes 1177240356 M * doener he's on vacation, so here's a work-around until daniel_hozac can fix it: http://people.linux-vserver.org/~doener/magic-umount.c 1177240378 M * doener call that from postpost-stop as "magic-umount /data" 1177240421 M * doener it's pretty hacky, but copes with the strange rbind semantics 1177240458 M * doener you'll need the "umount -n /vservers/test/data" and that one 1177240482 M * oliwel wtf is this.... 1177240502 M * oliwel looks like a new kind of chdir bomb ;) 1177240509 M * doener the "special" chdir call makes it hit the rbind'ed / 1177240528 M * doener I've just chosen enough ..'s to cope with most directory layouts 1177240534 M * oliwel *lol* 1177240536 M * doener didn't want to count how many you need 1177240542 M * oliwel will try - hang on 1177240642 M * oliwel no luck 1177240644 M * oliwel umount: No such file or directory 1177240654 M * oliwel still here /dev/drbd/test /data ext3 rw,data=ordered 0 0 1177240675 J * onox ~onox@kalfjeslab.demon.nl 1177240689 M * doener hm, maybe (the probably undefined) kernel semantics for the rbind to / changed? 1177240707 M * oliwel I am not a kernel guru... 1177240718 M * doener been talking to myself ;) 1177240719 M * oliwel I will try to load the old kernel and see what happens 1177240732 M * oliwel have to leave for an hour.....thx anyway till here... 1177240744 M * oliwel if you have an idea just post here - will be back in a good hour 1177240750 Q * boci^ Ping timeout: 480 seconds 1177240759 M * doener if the kernel semantics changed, that could probably break uv though 1177240791 M * oliwel you can also answer on the ML - there is a post from yesterday about this issue 1177240793 M * oliwel gotta go.... 1177240800 M * doener cya 1177240829 J * FireEgl Sebastian@adsl-147-76-85.bhm.bellsouth.net 1177240931 M * tanjix is there somewhere an older howto on how to apply the vserver patch to a vanilla kernel? 1177240986 M * derjohn tanjix, an older howto? or an older patch ? 1177241008 M * tanjix just the way on how to apply the vserver patch to the vanilla kernel 1177241017 M * tanjix something with patch -p1 or so 1177241019 Q * mstrobert_ Read error: Operation timed out 1177241070 M * tanjix trying to renew the debian etch kernel linux-image-2.6.18-4-vserver-686 to support ram up to 64gb 1177241088 M * tanjix and i think i must apply server patches 1177241108 M * tanjix the renew itself worked but testme.sh shows errors and i think that the kernel has no vserver functions in it 1177241153 Q * shedi Quit: Leaving 1177241274 M * derjohn tanjix, well, as Debian includes the vserver patches, you can get the source from debian and build the "flavor vserver" from it. 1177241305 M * daniel_hozac doener: you reach the rbind'ed / only by going just far enough to hit the real /, no? 1177241315 M * derjohn tanjix, as Debian changed the kernel build system recently, I cannot help you in that way, maybe "waldi" can. 1177241320 M * daniel_hozac IIRC one level above that hits the chroot'ed /. 1177241334 M * doener daniel_hozac: there's no chroot in postpost-stop, is there? 1177241338 M * daniel_hozac no. 1177241372 M * derjohn tanjix, I build my kernels from vanilla sources and apply newest vserver releases, debian currently lacks somewhat behind. 1177241399 M * tanjix derjohn: just need the command on how to apply the patch to the vanilla kernel 1177241435 M * derjohn tanjix, patch -p1 -i patch_you_want, maybe try with --dry-run before. 1177241466 M * tanjix must i run that inside the directory where the vanilla kernel is ? 1177241502 M * derjohn tanjix, but take care, the 2.6.18 needs an older patch ... you should rethink about going to 2.6.20 or ask harry about a prepatched 2.6.19 (i think he run machines with 16 GB RAM or such) 1177241515 M * derjohn tanjix, yes, it must be in the kernel src root dir 1177241527 M * doener daniel_hozac: anyway, the question is: why are the mounts still active in postpost-stop at all 1177241531 M * tanjix i am using kernel 2.6.19-7 and the patch from the website 1177241536 M * tanjix 89-149-195-244:/usr/src/linux-2.6.19-7-vserver-686# patch -p1 ../patch-2.6.19.7-vs2.2.0.diff 1177241542 M * derjohn tanjix, use make-kpkg to build .debs from you source install manuall install 1177241544 M * daniel_hozac doener: because vserver ... stop has to be executed inside the guest's namespace. 1177241544 M * tanjix that seems to do nothing? :( 1177241555 M * derjohn tanjix, open eyes ;) 1177241563 M * derjohn tanjix, patch -p1 -i patch_you_want, maybe try with --dry-run before. 1177241566 M * doener daniel_hozac: postpost-stop is run after umountVserver was called 1177241566 M * daniel_hozac doener: it tests for the presence of /etc/{init,rc}.d/rc 1177241575 M * derjohn wasnt there a "minus i" ? 1177241587 M * tanjix oh :) *cleaning glasses* hehe 1177241587 M * daniel_hozac doener: the rbind should still be active. 1177241594 M * derjohn os do "patch -p1 < ../patch-2.6.19.7-vs2.2.0.diff" 1177241613 M * tanjix ok patch applied, thank you derjohn! 1177241620 M * derjohn here you could get prepatches sources http://blog.derjohn.de/snipsnap/space/start/2007-04-22/1#My_(derjohn's)_Linux-VServer_Kernels_for_Debian_updated_to_new_Version 1177241640 M * doener daniel_hozac: oliwel's paste shows that all mounts, the original and the rbind'ed ones are still active 1177241644 M * derjohn errh ... http://blog.derjohn.de .. but nothing with 64G enabled.- 1177241648 M * doener it might break due to his manual mount though 1177241666 M * derjohn tanjix, have fun in (re)building :) 1177241719 M * daniel_hozac tanjix: did you already try using the amd64 kernel instead? 1177241738 M * tanjix daniel_hozac: yes, but i was unable to install amd64 distro, it failed on some things 1177241749 M * daniel_hozac such as? 1177241767 M * derjohn tanjix, you can use an amd64 kernel on a i386 userspace 1177241770 M * tanjix after installing from cd server could not boot up due to some I/O errors 1177241788 M * derjohn daniel_hozac, I know suc problems from acpi on tyan servers ... 1177241800 M * derjohn *such (but suck would also fit) 1177241801 M * daniel_hozac doener: umountVserver does nothing if using namespaces, if i'm reading the logic correctly. 1177241828 M * tanjix derjohn: i tried to use the config from /boot/config-2.6.18-4-vserver-686 to the new patched kernel. would that have success? 1177241844 M * daniel_hozac (the x || y syntax throws me every single time...) 1177241856 M * doener dito 1177241904 M * doener especially because of the inverted boolean meaning of 0 1177241913 M * daniel_hozac yeah. 1177241972 M * doener hm, did 0.30.210 have namespace support by default? 1177241977 M * daniel_hozac yep. 1177241992 M * daniel_hozac namespaces have been the default since 0.29.something, i think. 1177242018 M * doener wow 1177242018 M * daniel_hozac but the vserver ... stop entering the namespace was added in 0.30.212. 1177242046 M * daniel_hozac (so that guests with the root mount in the fstab and initstyle sysv could be stopped) 1177242072 M * doener and earlier, we had the namespace already dead when postpost-stop was running? 1177242103 Q * mire Quit: Leaving 1177242125 M * doener cause otherwise, I can't explain why oliwel's setup worked at all 1177242127 J * mstrobert ~mstrobert@wkstn.wycliffe.ca 1177242127 M * daniel_hozac yep. 1177242130 M * doener ok 1177242149 M * derjohn tanjix, well, there aer always some changes on .config from kernel to kernel .... 1177242166 M * doener daniel_hozac: so basically we lost the possibility to run scripts after the namespace is dead 1177242175 M * daniel_hozac indeed. 1177242190 M * derjohn tanjix, make oldconfig before ... 1177242191 M * daniel_hozac i guess running those particular checks inside the namespace would be better. 1177242213 M * tanjix derjohn: after i integrated the config from the 2.6.18-4 kernel? 1177242243 M * doener daniel_hozac: that would require some "keep-alive" process, right? 1177242252 M * daniel_hozac doener: hmm? 1177242268 M * derjohn yes.... if you apply the vs 2.2.0 there are even some new vserevr features which need to be en- or disabled. 1177242283 M * tanjix ok, i ran make oldconfig 1177242291 M * tanjix now i can start the compile? 1177242293 M * doener daniel_hozac: won't the namespace die off because there's noone using it? or does that happen late enough? 1177242307 M * derjohn tanjix, but you could try my amd64 kernel, as it is a 2.6.20 and may sata things are fixed there. 1177242319 M * tanjix is the ahci driver indluded? 1177242332 M * daniel_hozac doener: the checks i'm referring to decide whether to run /etc/init.d/rc or /etc/rc.d/rc to stop the guest. 1177242364 M * doener someone fluent in mutt? I once again forgot how to tag all message that were sent in a certain month, e.g. all from January 2007 1177242370 M * doener ah, ok 1177242609 J * baldy ~baldy@pptp.dial.ipv6-network.de 1177242614 M * tanjix someone could give me more machien power to speed up this compiling? :p 1177242809 M * doener nvm, found out 1177242830 M * derjohn tanjix, star-hosting has some vservers to rent, with some them it gets faster *g* 1177242839 A * derjohn is away for an hour ... c u ! 1177242872 M * derjohn tanjix, ahci should be in vanilla since 2.6.18 IIRC ... (vt8251 .... ) bye ! 1177245068 J * arachnist ~arachnist@088156189068.who.vectranet.pl 1177246105 M * tanjix anyone an idea what has gonewrong here? www.server.cd/problem.jpg 1177246131 M * tanjix this is the kernel i built what i asked before 1177246538 N * DoberMann DoberMann[PullA] 1177247503 Q * FireEgl Quit: BBL 1177247734 M * baldy tanjix: forget to setup the correct fs type? 1177247932 N * DoberMann[PullA] DoberMann 1177247995 M * derjohn tanjix, i bet you didnt update initrd or such. 1177248035 M * derjohn tanjix, it always a better way to build .deb packages from the kernel, as dpkg will do everything needed if you install that way. 1177248058 M * derjohn tanjix, google for make-kpkg and kernel-package 1177248808 Q * bzed Quit: Leaving 1177249468 M * tanjix derjohn: i did that 1177249479 M * tanjix i ran make-kpkh --initrd --revision v01 1177249488 M * tanjix and then i installed that .deb file 1177249634 M * tanjix make-kpkg --initrd --revision v01 kernel_image 1177249652 M * tanjix this is what i did to create the -.deb 1177250129 Q * bonbons Quit: Leaving 1177250280 Q * bragon_ Quit: Lost terminal 1177250292 M * tanjix i also tried with the 2.6.20 kernel and its patch, same error when rebooting :( 1177251066 J * boci^ ~boci@pool-1834.adsl.interware.hu 1177251385 M * tanjix ok i was able to solve the issue :) 1177251543 J * oliwel_ ~chatzilla@ppp-82-135-80-144.dynamic.mnet-online.de 1177251758 J * oliwel__ ~chatzilla@ppp-82-135-94-87.dynamic.mnet-online.de 1177251873 Q * oliwel Ping timeout: 480 seconds 1177252143 J * oliwel___ ~chatzilla@ppp-62-245-208-66.dynamic.mnet-online.de 1177252145 N * oliwel___ oliwel 1177252173 Q * oliwel_ Ping timeout: 480 seconds 1177252218 J * oliwel___ ~chatzilla@ppp-82-135-82-139.dynamic.mnet-online.de 1177252314 J * FireEgl Proteus@Sebastian.Atlantica.US 1177252328 Q * oliwel__ Ping timeout: 480 seconds 1177252648 Q * oliwel Ping timeout: 480 seconds 1177252839 M * derjohn tanjix, sorry was AFK, but nice to hear you solved it 1177253043 M * tanjix derjohn: no problem. i heard that some things changed from kernel 2.6.19 on 1177253050 M * tanjix for that my ahci drivers were not there any more 1177253061 M * tanjix i had to reactivate them and rebuilt the kernel and it worked 1177253153 M * oliwel___ daniel_hozac: doener: knocknock 1177253163 M * oliwel___ :Q 1177253203 M * daniel_hozac oliwel___: easier to just ask your question... 1177253230 M * oliwel___ daniel_hozac: I saw that doener discussed my problem with you - any ideas so far 1177253238 N * oliwel___ oliwel 1177253273 M * derjohn tanjix, you mean in vserver or in general? well, 2.6.20 came with many things, like kvm etc. which brought severe changes (that why some ppl stick to 2.6.19) 1177253276 M * daniel_hozac well, i guess we know what happens. 1177253314 M * daniel_hozac you'll have to run your postpost-stop script after vserver ... stop is finished. 1177253315 M * tanjix derjohn: i think the changed were general things - but everything works well now with 2.6.20.4 kernel 1177253320 M * derjohn tanjix, I can only tell you that I use my 2.6.20 vanilla pack in production and update etch server via apt get. and I am able to boot afterwards (with initrafs and root on md0 device) 1177253329 M * tanjix testme.sh shows everything with "success" 1177253354 M * derjohn .4 ? well, did oyu check changes in .5 .6 .7 ? 1177253367 M * tanjix no? 1177253401 M * oliwel daniel_hozac: ergh - so you mean "outside" the vserver-utils ? 1177253403 M * derjohn check kernel.org - or are you referring to 2.6.18-4 (debian) ? you can easily get incrementaö patches from .4 to .5 etc. 1177253413 M * daniel_hozac oliwel: yep, unfortunately. 1177253418 M * oliwel oor is there are postpostpostpost-stop script 1177253422 M * oliwel :( 1177253431 M * daniel_hozac oliwel: i'll see if i can get some sort of fix into 0.30.213. 1177253439 M * oliwel or my understanding - whats the problem and why worked it so far 1177253482 M * oliwel I rolled backed the urtils to the old version I used (.210) but it didnt work either - so I guess its a kernel issue not a tools one 1177253501 M * oliwel I will kick this keyboard out of service....loosing letters 1177253510 M * daniel_hozac it is a tool "issue". 1177253549 M * daniel_hozac in order for guests with the root mount in the guest's fstab and the sysv initstyle to work, some tests need to happen inside the namespace. 1177253555 M * daniel_hozac i fixed that in 0.30.212. 1177253559 M * oliwel hm - but why does a rollback to .210 bring the same error 1177253579 M * daniel_hozac vanilla 0.30.210? 1177253583 M * oliwel yep 1177253588 M * daniel_hozac it shouldn't. 1177253605 M * oliwel hm 1177253618 M * oliwel will try again 1177253650 M * oliwel do you have an idea how to fix it meanwhile ? anything with fancy namespace hacking or so 1177253984 M * oliwel daniel_hozac: crazy.... 1177253986 M * oliwel util-vserver: 0.30.210; Apr 22 2007, 16:55:33 1177254008 M * oliwel I cant see the mont in /proc/mounts but shuting down the drbd device fails 1177254022 M * daniel_hozac in postpost-stop? 1177254026 M * oliwel jep 1177254039 M * oliwel and know I get ugly kernel dumps on the drbd driver..... 1177254041 M * oliwel urgh 1177254053 J * dna ~naucki@185-225-dsl.kielnet.net 1177254059 M * oliwel Fixing recursive fault but reboot is needed! 1177254063 M * oliwel that doesnt look good 1177254073 M * daniel_hozac i'd put my money on that's why you can't shut it down. 1177254104 J * SoftIce ~phil@vc-196-207-45-253.3g.vodacom.co.za 1177254107 M * SoftIce hello :) 1177254160 M * SoftIce daniel_hozac leet, please tell me if I have an ipv6 host, what is the easiest way to trace it back to its ipv4 host ? 1177254199 M * oliwel daniel_hozac: you will loose - I can shutdown the disk when I issue the shutdown commadn after vserver 1177254225 J * shedi ~siggi@ftth-237-144.hive.is 1177254247 M * daniel_hozac oliwel: and how does that rule out that the dump is why you can't shut it down from the postpost script? 1177254299 M * daniel_hozac oliwel: you'd want to get that fixed either way. 1177254304 M * oliwel the device complains that there is an active mount 1177254319 M * oliwel daniel_hozac: I must see what causes the dump 1177254338 M * daniel_hozac SoftIce: what? you want to get an IPv4 address from an IPv6 address? 1177254404 M * SoftIce daniel_hozac: well ipv6 tunnel, i have an ipv6 address, its with bt exact. but how can I find out what ipv4 address it links to ? 1177254743 M * sid3windr you can't 1177254767 M * sid3windr they're totally separate protocols 1177254774 M * sid3windr you could of course hax0r the bt exact tunnel broker server 1177254792 M * SoftIce sid3windr: so no way to find out the ipv4 part of it? 1177254796 M * sid3windr no 1177254802 M * sid3windr there is no "ipv4 part" of an ipv6 address 1177254805 M * SoftIce no lookup techniques, etc 1177254818 M * SoftIce sid3windr: yes, i understand that, hard to explain but i think you understand 1177254823 M * SoftIce something has to initiate a tunnel 1177254834 M * SoftIce the ipv4 address would bind the ipv6 host 1177254836 M * sid3windr I don't have a tunnel, but I have ipv6 1177254850 M * SoftIce yes, thats a difference 1177254853 M * SoftIce as thats assigned to you 1177254867 M * sid3windr :) 1177254927 M * oliwel daniel_hozac: ok - using the old (.210) tools I get a coredump from drbd, usnig the 212 toolset the unmount fails because the device is busy - but no dump occurs 1177254956 M * oliwel I can umount the device after vserver stop has finished by hand - without dumping it 1177254965 M * SoftIce if using a 2001 you could have for instance packets to it ethernet(ipv4(ipv6(tcp(irc)))) 1177254987 M * oliwel if you are interessted in the dump see http://paste.linux-vserver.org/1546 1177255004 A * oliwel is out for a coffee 1177255221 M * SoftIce sorry 2022 1177255222 M * SoftIce 2002 1177255972 M * Nam question... when assigning IP's to a guest, are they added as a single interface per IP, or can like iproute2 and a single interface with multiple IP's assigned to that interface? 1177256021 M * daniel_hozac Nam: yes. 1177256046 M * daniel_hozac Nam: i.e. that's up to you. if you set the name for an interface, the utils will create an alias. 1177256059 M * daniel_hozac if not, they'll just add the address to the interface. 1177256156 M * Nam ok... so if you have a nodev file, it's add the specified interface for the vserver to the default "eth0" and so long as I don't specify the dev file, each ip is added via the iproute2 method? 1177256175 M * Nam or can I specify multiple ip addresses under in interface entry? 1177256193 M * Nam /etc/vservers/guest/interfaces/0/ip 1177256222 M * daniel_hozac uh? 1177256230 M * daniel_hozac with nodev, the utils won't touch the IP addresses. 1177256237 M * daniel_hozac nodev means that you've already created them. 1177256289 M * Nam is the interface in the guest the same as the host? or is it a virtual one inside the guest? 1177256310 M * daniel_hozac same as the host. 1177256334 M * Nam ok, so if I have nodev set, and I have an ip on eth2, it will show up as eth2 in the guest right? 1177256340 M * daniel_hozac yep. 1177256420 M * Nam ok... say I specify ip1 in /etc/vservers/guest/interfaces/0/ip and then I specify ip2 in /etc/vservers/guest/interfaces/1/ip... if those ip's are both setup with nodev, and they both exist on the same interface, they will appear in the guest as the same interface correct? 1177256433 M * daniel_hozac yes. 1177256495 M * Nam no other ip's will be visible right, other then the ones I specify correct?... i know before I was still seeing all the ip's, but that was the legacy version 1177256514 M * daniel_hozac exactly. 1177256571 M * Nam any ip that is being setup can be accessed by both the guest and host right?... if ip1 was setup on the guest, and the host had an application which already was binding to a port, that would conflict right? 1177256594 M * daniel_hozac yup. 1177256617 M * Nam gotta love when all the answers are "yup, yes, and exactly" :) thanks 1177256631 M * Nam you and Bertl have been a great help 1177256636 M * daniel_hozac heh, you're welcome. 1177256642 M * SoftIce anyone know what happend to bugtraq? 1177256664 M * Nam I haven't talked with harry yet, but if I recall, he's been the one going the gentoo work, so he has been of great help as well 1177256672 M * Nam ;) 1177256703 M * daniel_hozac hmm? harry handles the grsec+vserver merge. 1177256718 M * daniel_hozac Hollow and phreak`` are the Gentoo maintainers. 1177256726 M * Nam ah, right 1177256737 M * Nam all of the above have helped, hehe 1177256756 M * Nam i use grsec+vserver on one of our current servers 1177256767 M * SoftIce daniel_hozac: does Hollow and phreak have generic Gentoo kernels? 1177256772 M * Nam i'm making our new version which is going to hopefully go public soon though 1177256775 M * daniel_hozac SoftIce: what? 1177256778 M * SoftIce eg: similar to the 1 on the livecd but with grsec +vserver? 1177256789 M * Hollow we don't have ebuilds for vserver+grsec 1177256807 M * Nam any idea on when those are going to be coming along? 1177256810 M * phreak`` SoftIce: we have an official kernel, thats sys-kernel/vserver-sources, and I have an somewhat unofficial kernel, vsgrsec-sources in my overlay 1177256816 M * Hollow but it should not be a big deal to patch it yourself .. 1177256836 M * SoftIce overlay? 1177256844 M * SoftIce i'd like to try that vsgrec-source :) 1177256857 M * Nam are you going to be putting it into the public tree anytime soon? 1177256861 M * phreak`` SoftIce: yeah, overlay :P 1177256863 M * phreak`` Nam: no 1177256875 M * Nam you ever planning on it? 1177256889 M * SoftIce phreak``: that vsgresec is that a generic kernel, or would 1 have to still configure devices, etc? 1177256891 M * Nam or is it only going to be the vserver which is publicly released in the gentoo tree? 1177256936 M * phreak`` Nam: don't think so yet, as it really is overlay material ... basically because its a "mixed" together patch (no pun intended towards Harry or Spender, both do a great job) 1177256945 Q * dna Ping timeout: 480 seconds 1177256958 M * phreak`` SoftIce: eh ? you'll have to create your own kernel config, yes 1177256973 M * onox vsgrsec-sources \o/ 1177256989 M * SoftIce phreak``: then its just as easy to get vserver-source and patch with grsec ? 1177257023 M * Nam i think it would be easier to get gentoo-sources and then patch with the vserver/grsec patch 1177257035 M * SoftIce phreak``: what does this mean no pun intended towards Harry or Spender, both do a great job ? 1177257035 M * onox phreak``: vsgrsec-sources is hardened-sources + vserver patch? 1177257058 M * phreak`` onox: hardened-sources minus the grsec patch + harry's grsec vserver patch 1177257069 J * dna ~naucki@185-225-dsl.kielnet.net 1177257073 M * phreak`` SoftIce: I don't mean to disrespect their work :) 1177257083 M * SoftIce I see 1177257090 M * SoftIce and you are a developer with gentoo ? 1177257101 M * phreak`` yes, I am 1177257104 Q * meandtheshel1 Quit: Leaving. 1177257133 M * SoftIce phreak``: well whats the link to your ~user so I can get that kernel ? 1177257137 M * Nam and thank you for that work phreak`` 1177257146 M * SoftIce :) 1177257177 M * onox vsgrsec is funny: people asking why they only see sshd + their shell :p 1177257235 M * phreak`` SoftIce: http://overlays.gentoo.org/dev/phreak/browser/hardened/, but you probably want to use svn to get that branch .. (svn co http://overlays.gentoo.org/svn/dev/phreak/hardened/) 1177257268 M * SoftIce great thanks 1177257353 J * meandtheshell ~markus@85-125-193-29.dynamic.xdsl-line.inode.at 1177257607 M * SoftIce phreak``: your kernel has it got all the basics that gentoo requires, eg: prompt for dev, file system - /proc filesystem, virtual memory file system, etc etc 1177257616 M * SoftIce are they all allready checked in your config? 1177257647 M * phreak`` SoftIce: no 1177257658 M * phreak`` SoftIce: the are probably guessed from your current kernel config 1177257672 M * SoftIce ahh, I see. thanks 1177257726 M * SoftIce you dont have your .config posted publicy, a nicely selected kernel with the correct grsec config as not to cause issues within vservers? 1177257752 M * phreak`` I do, iirc 1177257770 M * phreak`` SoftIce: http://dev.gentoo.org/~phreak/configs/epimetheus/kconfig/vsgrsec-sources-2.6.19-r1.config 1177257801 M * phreak`` SoftIce: but bare in mind, that this is *very* hardware specific (as in just my raid controller, ide controller and stuff I *Really* need for that box) 1177257846 M * SoftIce great, thanks alot 1177257847 M * SoftIce :) 1177259304 J * kwowt ~quote@pomoc.ircnet.com 1177259313 M * kwowt derjohn, here? 1177259481 J * bonbons ~bonbons@83.222.38.145 1177259529 M * kwowt anyone using his kernel patch for 256IPs 1177259541 M * daniel_hozac why do you ask? 1177259558 M * kwowt cuz i'm a newbie and i have some questions:) 1177259613 M * onox kwowt: create a for loop and build 256 vservers? 1177259640 M * kwowt huh? 1177259645 M * daniel_hozac kwowt: so, ask. 1177259648 M * sid3windr onox: eh? :) 1177259661 M * kwowt i just apply the patch to network.h and then recompile the kernel, right? 1177259667 M * daniel_hozac yes. 1177259683 M * onox sid3windr: ;) 1177259737 M * kwowt cool 1177259745 M * kwowt no changes in kernel config needed or anything? 1177259756 M * daniel_hozac no. 1177260249 Q * id23 Ping timeout: 480 seconds 1177260364 J * Piet hiddenserv@tor.noreply.org 1177260466 Q * arachnist Quit: bbl 1177263143 M * tanjix vps023:~# vserver vs5322 start 1177263143 M * tanjix ncontext: vc_net_create(): Invalid argument 1177263148 M * tanjix what's wrong? :( 1177263891 M * daniel_hozac you disabled dynamic contexts and didn't specify a static one for that guest. 1177264045 M * tanjix hm is this a change in vs2.3? 1177264074 M * tanjix how can i enable dynamic contexts? 1177264091 M * daniel_hozac in the kernel config. 1177264098 M * daniel_hozac but you really should use static context ids. 1177264148 M * tanjix hm maybe i will do that later, but for now i have to move more than 2.000 vservers to new hardware 1177264160 M * tanjix do you know where in the kernel config i must do the changes? 1177264217 M * daniel_hozac your guests already seem to have numeric ids. 1177264222 M * daniel_hozac why not use that as the id? 1177264235 M * tanjix because there are letters in it 1177264244 M * daniel_hozac so, trim the vs. 1177264261 M * tanjix and it would be too much work te reconfigure all software parts from our tools for this short time 1177264297 M * tanjix if a customer reinstalles the server, the context is away, so for now i must have them dynamically allocated 1177264436 M * tanjix where in the kernel config can i find dynamic context ? 1177264472 Q * shedi Quit: Leaving 1177264489 M * daniel_hozac somewhere in the Linux-VServer menu. 1177264495 M * daniel_hozac you have to enable legacy for it to show up at all. 1177264551 M * tanjix i don't see the one or the other? 1177264567 M * tanjix no legacy, no dynamic context 1177264591 M * daniel_hozac well, what kernel do you have? 1177264596 M * tanjix 2.6.20.4 1177264612 M * daniel_hozac and vserver version? 1177264613 J * tzafrir_laptop ~tzafrir@bzq-88-153-210-69.red.bezeqint.net 1177264621 M * tanjix 2.3.xx 1177264637 M * tanjix 2.3.0.12 1177264639 M * daniel_hozac you know that's a development version, right? 1177264650 M * daniel_hozac i think all of the legacy cruft has been dropped from 2.3. 1177264716 M * tanjix what kernel should i build then? 1177264737 M * daniel_hozac latest stable seems like a good choice, no? 1177264766 M * onox is this about vserver create? 1177264800 M * tanjix daniel_hozac: :p sure 1177264831 M * onox then create a wrapper, that assigns a context id? 1177264858 M * daniel_hozac vserver ... build in 0.30.213+ already assigns static ids by default. 1177265306 M * kwowt Hm 1177265325 M * kwowt i'm getting the error "vshelper.init: can not determine xid of vserver 'login'; returned value was ''" 1177265334 M * kwowt cant find anything usefull on google 1177265342 M * kwowt this happens when i try to start the vserver 1177265348 M * kwowt the guest vserver 1177265350 M * daniel_hozac kwowt: your guest isn't starting any process, so the guest dies before it's done starting it. 1177265368 M * kwowt which means ...? 1177265381 M * daniel_hozac that you should make it start some sort of service. 1177265398 M * kwowt i dont get it 1177265400 J * independence independen@blinkenshell.org 1177265429 M * independence I'm having some problems with user-level quota inside guest host 1177265445 M * independence It used to work fine for a long time, but since some reboot/upgrade I can't do quotaon anymore: 1177265449 M * independence titan / # quotaon / 1177265449 M * kwowt .. 1177265451 M * independence quotaon: Mountpoint (or device) / not found. 1177265454 M * independence titan / # quotaon /dev/hdv1 1177265456 M * independence quotaon: using //aquota.group on /dev/hdv1 [/]: No such device 1177265459 M * independence quotaon: using //aquota.user on /dev/hdv1 [/]: No such device 1177265467 M * kwowt independence wait for ur turn ;p 1177265469 J * ema ~ema@rtfm.galliera.it 1177265482 M * daniel_hozac independence: does /dev/hdv1 exist? 1177265491 M * independence daniel_hozac: yes 1177265492 M * daniel_hozac independence: is it a vroot device? 1177265499 M * independence titan / # mount 1177265499 M * independence /dev/hdv1 on / type ufs (rw,acl,usrquota,grpquota,user_xattr) 1177265503 M * daniel_hozac independence: have you associated it with the correct device? 1177265515 M * daniel_hozac independence: have you loaded the vroot module? 1177265560 M * independence I have done vrsetup /dev/vroot1 /dev/vg1/titan_root 1177265562 M * independence etc 1177265595 M * kwowt .. 1177265675 M * daniel_hozac independence: and /dev/vroot1 is the same as /dev/hdv1 inside the guest? 1177265732 M * independence hm, /dev/vroot1 is not visible inside the guest 1177265741 M * daniel_hozac it shouldn't be. 1177265751 M * daniel_hozac it should be called /dev/hdv1 inside. 1177265761 M * independence yes, in the guest I can see /dev/hdv1 1177265769 M * daniel_hozac but is it the same as /dev/vroot1? 1177265802 M * independence How do I check that? 1177265813 M * daniel_hozac ls -l /dev/vroot1 /vservers//dev/hdv1 1177265829 M * SoftIce hehe, can tell you creating this dam tar.bz2 file of my 3 vservers is taking for ever, been running for a few hours now, and its only used about 33gigs :) 1177265850 M * daniel_hozac SoftIce: i assume that's all data, right? 1177265852 M * independence saturn independence # ls -l /dev/vroot1 /vservers/titan/dev/hdv1 1177265852 M * independence brw-rw---- 1 root disk 4, 1 2007-04-20 16:46 /dev/vroot1 1177265852 M * independence brw-rw---- 1 root disk 4, 1 2006-11-26 18:55 /vservers/titan/dev/hdv1 1177265855 M * onox kwowt: which distro do you run? 1177265885 M * daniel_hozac independence: so, that looks fine. what kernel are you running? 1177265899 M * independence 2.6.19.2-grsec2.1.10-vs2.2.0-rc8 1177266013 M * daniel_hozac you might want to upgrade to something recent, but i don't recall us fixing anything related to quota recently. 1177266067 J * xe ex@valis.net.pl 1177266067 Q * xe 1177266272 M * kwowt * Starting lo 1177266272 M * kwowt * network interface lo does not exist 1177266272 M * kwowt * Please verify hardware or kernel module (driver) 1177266282 M * kwowt this is what happens when i try to start sshd in guest vserver .. 1177266285 M * kwowt any idea about this :p? 1177266313 M * onox check if net.lo exists in /etc/init.d? 1177266327 M * onox kwowt: and how do you get into a vserver if it doesn't start? :| 1177266342 M * kwowt it exists 1177266354 M * kwowt lo is up in root server 1177266360 M * kwowt ist just doesnt wanna start in guest vserver 1177266361 M * onox < kwowt> i'm getting the error "vshelper.init: can not determine xid of vserver 'login'; returned value was ''" 1177266384 M * kwowt nevermind that 1177266398 M * onox yes, I mind :p 1177266403 M * onox how do you get into the vserver? 1177266416 M * onox I got that error too, and I had to use a bloated stage4 instead of a stage3 1177266422 M * kwowt i just used a plain initstyle :> 1177266433 M * onox kwowt: what distro? 1177266440 M * kwowt OS? 1177266442 M * kwowt gentoo 1177266500 M * onox k 1177266512 M * onox did you use stage3? 1177266515 M * kwowt yes 1177266525 M * onox and created the vserver with plain init? 1177266527 M * onox not gentooinit? 1177266532 M * kwowt yea 1177266536 M * onox hmm 1177266575 M * onox build your own stage4 with catalyst, or use a pre-build stage4 and start unmerging bloat 1177266605 M * kwowt bloat? 1177266641 M * onox yes, any pkg that was installed by someone else and you don't like :p 1177266656 M * kwowt :) 1177266658 M * onox ;) 1177266673 M * kwowt last week i found some stage3 which didnt make me any problems 1177266675 M * kwowt but i forgot where 1177266680 M * onox hmm 1177266722 M * onox what's stupid, Gentoo says I have to get into the vserver somehow to start add a service, but vserver doesn't allow me to start it, and chrooting didn't help either 1177266809 M * kwowt :) 1177266917 M * onox daniel_hozac: can't you just keep the guest living even when it doesn't start any service? 1177266955 M * kwowt * network interface lo does not exist 1177266957 M * kwowt argh! 1177267186 M * onox kwowt: kwowt do you have a symlink net.lo? 1177267235 J * mire ~mire@38-169-222-85.adsl.verat.net 1177267476 M * PowerKe the rescue script just runs 'sleep' to keep it running, so I guess you could use that temporarily. I suppose that in the end you'll be adding services anyway. 1177267532 M * kwowt lrwxrwxrwx 1 root root 29 Jan 13 19:35 net.lo -> ../../lib/rcscripts/sh/net.sh 1177267621 M * PowerKe adding a service should work with chroot /vservers/yourguest and then rc-update add default 1177267888 M * tanjix vps023:~# vserver vs5102 start 1177267888 M * tanjix Failed to start vserver 'vs5102' 1177267888 M * tanjix vps023:~# 1177267894 M * tanjix where can i see the reason? 1177269526 M * daniel_hozac tanjix: that's all you get?! 1177269574 M * daniel_hozac onox: well, you could make it persistent. 1177269590 M * daniel_hozac (if your kernel supports it, but that's 2.0.2+) 1177270012 M * SoftIce what happend to bugtraq? 1177270015 M * SoftIce why did they close it? 1177270085 Q * infowolfe Ping timeout: 480 seconds 1177270272 M * derjohn kwowt, I think you issue was solved? 1177270642 M * kwowt wasnt 1177270645 M * tanjix daniel_hozac: yes that's all 1177270649 M * kwowt still cant run net.lo 1177270665 M * daniel_hozac kwowt: what stageX are you using? 1177270668 M * kwowt 3 1177270668 M * oliwel daniel_hozac: hi - I try to hack around the start/stop scripts and fail on one question -. how can I leave a vnamespace ? 1177270670 M * daniel_hozac kwowt: Hollow's stage4? 1177270680 M * kwowt umm 1177270683 M * kwowt i think it was 1177270687 M * kwowt sec 1177270690 M * daniel_hozac oliwel: you can't. 1177270696 M * oliwel :( 1177270704 M * oliwel shit 1177270706 M * kwowt yes,it was stage4 now 1177270734 M * oliwel daniel_hozac: how big is the chance you get a fix in 213 and when will this be ? 1177270739 M * daniel_hozac kwowt: stage4-x86-20070321.tar.bz2? 1177270778 M * kwowt yep 1177270797 M * daniel_hozac oliwel: you could just remove the stop part of the test on line 200 of /usr/sbin/vserver. 1177270846 M * oliwel ergh - but the server wont stop in this case? 1177270867 M * daniel_hozac if you have the root mount in the guest's fstab and you're using the sysv initstyle, right. 1177270896 M * oliwel I tried to add the drbd shutdown here but the script uses the namespace of the guest 1177270903 M * oliwel thats not what I expect 1177270922 M * oliwel I need a reliabel way to shutdwn the device with the server to make my failover work 1177270966 M * onox kwowt: try to run stage4 1177271040 M * PowerKe kwowt: Did you look at http://people.linux-vserver.org/~hollow/stages/ for the stage3 guest? (linked from http://www.gentoo.org/proj/en/vps/vserver-howto.xml) 1177271056 M * onox phreak``: alive? 1177271064 M * phreak`` onox: yes 1177271081 M * onox can I use basel-2 in a vserver guest too? 1177271107 M * phreak`` onox: not yet, since it has a "few" bugs .. 1177271116 M * kwowt i tried both 1177271118 M * onox what bugs? 1177271120 M * kwowt stage3 and stage4 1177271120 M * phreak`` onox: you can, if you fix those bugs apparently :) 1177271123 M * kwowt none runs net.lo 1177271132 M * onox lol, C is not my language 1177271155 M * onox kwowt: followed gentoo docs? 1177271170 M * PowerKe Why do you want to run net.lo? 1177271173 M * phreak`` onox: trying to modify some sysctl values (which is forbidden) and trying to access dmesg (which is forbidden from within VPS's too) 1177271185 M * kwowt i want to run sshd, but it automaticly wants to start net.lo 1177271195 M * onox phreak``: b-1 does the same thing too 1177271207 M * onox grsecurity complained about it 1177271214 M * PowerKe hmm, gonna try that 1177271215 M * daniel_hozac oliwel: you should remove the stop part of that statement, that should make the postpost-stop script work again. 1177271255 Q * SoftIce 1177271331 M * kwowt http://rafb.net/p/Yvnv7f24.html 1177271373 M * daniel_hozac kwowt: looks like a bug to me. 1177271385 M * kwowt whose bug 1177271394 M * daniel_hozac i don't see why sshd would require lo specifically. 1177271402 M * PowerKe got same problem... 1177271412 M * kwowt ... 1177271424 M * PowerKe I guess easiest would be to remove 'depend net' from the sshd script, but it used to work with that line 1177271434 M * daniel_hozac it just has depend net? 1177271445 M * daniel_hozac you guys _did_ build your guests with -d gentoo, right? 1177271454 M * PowerKe I did 1177271456 M * onox uhh 1177271460 M * onox -d is init style? 1177271460 M * oliwel daniel_hozac: you mean I should NOT run vserver.stop ? 1177271464 M * onox in that case: yes 1177271467 M * daniel_hozac oliwel: you should. 1177271474 M * oliwel ??? 1177271476 M * daniel_hozac onox: no, -d is distribution. 1177271483 M * onox oh, still yes :) 1177271486 M * kwowt i did too 1177271493 M * oliwel Line 200 in /usr/sbin/verver calls the vserver.stop scrpt 1177271494 M * onox running hardened + vserver server 1177271529 M * daniel_hozac oliwel: then you have a really old version. 1177271548 M * daniel_hozac line 200 is test "$2" != enter -a "$2" != stop || \ 1177271599 M * oliwel hm, than I guess I screwed up the system while switching the tools 1177271630 M * PowerKe hmm, the stage3 seems to use baselayout instead of baselayout-vserver 1177271642 M * daniel_hozac on Hollow's people page? 1177271666 M * oliwel daniel_hozac: I have this on line 204 - so I guess this is ok 1177271673 M * PowerKe daniel_hozac: yes 1177271685 M * daniel_hozac that's expected, baselayout 1.13+ is supposed to be able to run inside guests without problems. 1177271711 M * daniel_hozac (still some hickups though) 1177271721 M * kwowt :S 1177271740 M * daniel_hozac i think Hollow has made all of the necessary changes in util-vserver trunk though. 1177271779 M * PowerKe util-vserver: 0.30.212; Apr 20 2007, 20:53:25 1177271798 M * oliwel daniel_hozac: great.....now I get a core dump 1177271807 M * oliwel and a killed guest 1177271825 M * daniel_hozac oliwel: and a kernel dump, i presume? 1177271828 M * kwowt 0.30.212-r2 1177271838 M * onox kwowt: on the gentoo.org site are some explicit instructions how to create a vserver 1177271843 M * onox did you follow these? 1177271845 M * oliwel daniel_hozac: BUG: unable to handle kernel NULL pointer dereference at virtual address 00000030 1177271868 M * oliwel this is same as running the 210 version of the tools 1177271873 M * daniel_hozac oliwel: right. 1177271879 M * daniel_hozac oliwel: sounds like your drbd is broken. 1177271882 M * oliwel daniel_hozac: is this a bug in the drbd code ? 1177271887 M * oliwel daniel_hozac: shit 1177271894 M * kwowt i think i did, onox 1177271902 M * oliwel will try to go back on the one I used before 1177271905 M * kwowt last time i did the same thing, and it worked 1177271949 M * PowerKe hmm, just noticed /etc/init.d/util-vserver doesn't exist 1177271969 M * PowerKe will try to update to 213 first 1177272032 M * onox vserver foobar build --context 1 --hostname foobar --interface eth0:10.0.0.1/24 --initstyle gentoo -m template -- -d gentoo -t /root/stage4-amd64-20070322.tar.bz2 1177272038 M * onox that's what I did 1177272039 M * daniel_hozac onox: --context 1 is wrong. 1177272049 M * onox ah, 1 was some other number 1177272053 M * daniel_hozac it should be between 2 and 49151. 1177272057 M * kwowt ok 1177272062 M * onox vserver foobar build --context x --hostname foobar --interface eth0:10.0.0.1/24 --initstyle gentoo -m template -- -d gentoo -t /root/stage4-amd64-20070322.tar.bz2 1177272066 M * onox :p 1177272170 M * PowerKe onox: did you install it recently? 1177272200 Q * dev-0 Read error: Connection reset by peer 1177272216 M * onox PowerKe: yes, 2 weeks ago or something 1177272244 M * PowerKe what's your util-vserver version? 1177272294 M * onox PowerKe: uhh 1177272298 M * onox trunk? 1177272317 M * PowerKe vserver-info should tell :) 1177272322 M * daniel_hozac it doesn't. 1177272333 M * PowerKe It does here :) 1177272352 M * daniel_hozac so you modify the configure.ac when you build from trunk? 1177272363 M * PowerKe I just use the gentoo packages 1177272409 M * onox daniel_hozac: what's current revision? 1177272420 M * daniel_hozac 2533. 1177272459 M * onox is rev 2525 about 3 weeks old? 1177272476 M * daniel_hozac 04/08/07 1177272533 M * onox ok 1177272549 M * onox then I used rev 2525 +- 2 1177272887 M * matti Favorites sys-admin hours: 1:28, 2:56, 5:12, 10:24, 20:48. 1177272900 M * onox ? 1177272901 M * sid3windr you forgot 13:37 1177272903 M * onox lol :) 1177272914 M * matti Damn. 1177272921 M * onox and 01:01 1177272927 M * matti sid3windr: You're the third person who pointed this out. 1177272930 M * matti And... 1177272931 M * matti 4:03 1177272932 M * sid3windr :> 1177272934 M * matti 4:04 1177272936 M * matti Etc ;p 1177272939 M * sid3windr 4:02 \o/ 1177272944 M * onox 4:02? 1177272952 M * sid3windr yes, 402. the best HTTP status code :> 1177272953 M * matti 42 1177272958 M * onox haahah :D 1177272960 M * matti ;p 1177273003 M * onox 02:18 1177273023 M * onox hint: some desktop thingy :p 1177273697 M * PowerKe Hmm, just re-read Hollow's page, upgraded to util-vserver .213, still no luck with the net script 1177273828 M * daniel_hozac what does /etc/init.d/net contain? 1177273902 M * daniel_hozac ah, hmm. 1177273903 M * PowerKe net.lo, and it's long 1177273936 M * daniel_hozac try removing everything /etc/init.d/net.*, except net.vserver. 1177274239 M * PowerKe That might be it, just ran into something else now 1177274256 M * PowerKe btw, net.vserver is only created when creating the guest with util-vserver .213 not with .212 1177274289 M * daniel_hozac as expected, as it was added there... 1177274357 M * PowerKe so kwowt will at least have to upgrade to .213 as well 1177274397 Q * Piet Remote host closed the connection 1177274435 J * Piet hiddenserv@tor.noreply.org 1177274814 M * oliwel daniel_hozac: FYI - send a help to the drbd list - will keep you up to date ;) 1177274908 M * PowerKe It doesn't complain about the net anymore, just start-stop-daemon: open pidfile /var/run/sshd.pid: No such file or directory (starting sshd) 1177275041 M * PowerKe You need to do chroot /vservers/ and rc-update add net.vserver boot, to have it use the net.vserver script 1177275435 M * daniel_hozac Hollow? 1177275519 Q * ema Quit: leaving 1177275585 Q * bonbons Quit: Leaving 1177276114 M * PowerKe Works now... noob mistake, had sshd on the host bound to 0.0.0.0 1177276169 M * PowerKe kwowt: install util-vserver .213, recreate the guest, make sure the host sshd is not bound to the guest IP, add net.vserver to boot runlevel 1177276352 Q * oliwel Quit: ChatZilla 0.9.78.1 [Firefox 2.0.0.3/2007030916] 1177276599 M * PowerKe Hollow: the vserver-howto page still uses .212 (emerge >=sys-cluster/util-vserver-0.30.212) and doesn't mention net.vserver yet. 1177276652 J * AiRBuS-A380 8.2.6.65@88.226.162.191 1177276683 J * KraL ~aSi@88.246.28.241 1177276696 M * AiRBuS-A380 weasel 1177276698 T * KraL koim bikere 1177276701 M * AiRBuS-A380 aysel gibim olmusun 1177276702 M * AiRBuS-A380 (: 1177276704 M * KraL ;DDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDD 1177276704 M * KraL ;DDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDD 1177276706 M * KraL ;DDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDD 1177276707 M * AiRBuS-A380 Hahaha hAAAhahahahaha 1177276707 M * AiRBuS-A380 ¦D 1177276708 M * KraL ;DDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDD 1177276708 Q * KraL Killed (FloodServ ((FloodServ) Warning, you have triggered a network protection. Stop flooding!)) 1177276733 M * daniel_hozac PowerKe: because it's still really new. 1177276736 J * KraL ~aSi@88.246.28.241 1177276742 J * Aiken ~james@121.45.222.137 1177276742 F * ChanServ +o daniel_hozac 1177276743 M * KraL topic im hala burda 1177276744 M * KraL ;D 1177276747 M * AiRBuS-A380 ¦D 1177276760 M * KraL fotorafını çekmem lazım 1177276761 M * KraL ;D 1177276764 F * daniel_hozac +b *!*@88.246.28.241 1177276776 F * daniel_hozac +b *!*@88.226.162.191 1177276783 K KraL daniel_hozac . 1177276789 K AiRBuS-A380 daniel_hozac . 1177276831 T * daniel_hozac http://linux-vserver.org/ | latest stable 2.2.0, 2.0.3-rc2, devel 2.3.0.12, stable+grsec 2.0.2.1, 2.2.0 | util-vserver-0.30.212, 0.30.213-rc6 | libvserver-1.0.2 & vserver-utils-1.0.3 | He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the Wiki, and we'll forget about the minute ;) 1177276838 M * PowerKe daniel_hozac: No problem, but without it the stage3 doesn't work very well as soon as you try to start network services 1177276839 F * daniel_hozac -o daniel_hozac 1177276870 M * sid3windr mmkay 1177276872 M * sid3windr .tr ofcourse :p 1177276874 M * daniel_hozac PowerKe: yes, that's a really new stage3 too. 1177276879 M * daniel_hozac sid3windr: obviously. 1177277613 Q * baldy Ping timeout: 480 seconds 1177278082 N * DoberMann DoberMann[ZZZzzz] 1177278521 Q * Greek0 Server closed connection 1177278537 J * Greek0 ~greek0@85.255.145.201 1177278692 Q * ex Server closed connection 1177278700 J * ex ex@valis.net.pl 1177279458 Q * tanjix Ping timeout: 480 seconds 1177280696 J * baldy ~baldy@pptp.dial.ipv6-network.de 1177281079 Q * boci^ Quit: Távozom 1177281746 Q * Piet Quit: Piet 1177282027 Q * dna Quit: Verlassend 1177282263 Q * meandtheshell Quit: Leaving. 1177282377 J * varzamurata ~ss@bas2-montreal02-1096603617.dsl.bell.ca 1177282576 J * infowolfe ~infowolfe@c-67-164-195-129.hsd1.ut.comcast.net 1177282583 M * varzamurata www.Queeq.com Online market 100% Free !!! ; Marche en lignie 100% Gratuit !!! ; Piata online 100% Gratuita 1177282646 Q * micah Remote host closed the connection 1177282775 J * tanjix ~tanjix@office.star-hosting.de 1177282992 J * infowolfe_ ~infowolfe@c-67-164-195-129.hsd1.ut.comcast.net 1177282993 Q * infowolfe Read error: Connection reset by peer 1177283131 J * micah ~micah@micah.riseup.net 1177285964 Q * RungeKutta Ping timeout: 480 seconds 1177286182 Q * sc0tt Server closed connection 1177286195 J * sc0tt ~scott@209.51.169.84