1206662623 J * hparker ~hparker@208.188.248.81 1206662652 Q * mick_work Remote host closed the connection 1206663262 Q * hparker Remote host closed the connection 1206663306 J * hparker ~hparker@208.188.248.81 1206663354 Q * nebuchadnezzar Read error: Connection reset by peer 1206663370 J * nebuchadnezzar ~nebu@zion.asgardr.info 1206663791 Q * hparker Ping timeout: 480 seconds 1206663961 J * hparker ~hparker@208.188.248.81 1206665125 J * chasep ~chase@c-98-201-23-168.hsd1.tx.comcast.net 1206665343 Q * chasep 1206666056 Q * julius_ Remote host closed the connection 1206667061 Q * hparker Ping timeout: 480 seconds 1206668541 N * Bertl_zZ Bertl 1206668567 M * Bertl back again ... couldn't sleep 1206672265 Q * rgl Read error: Connection reset by peer 1206672712 Q * dowdle Remote host closed the connection 1206673627 Q * docelic Quit: http://www.spinlocksolutions.com/ 1206678088 Q * Hollow Remote host closed the connection 1206678103 J * Hollow ~hollow@proteus.croup.de 1206678238 Q * Hollow Remote host closed the connection 1206678251 J * Hollow ~hollow@proteus.croup.de 1206679242 Q * bronson Quit: Ex-Chat 1206679673 M * etteyafed Bertl: PING 1206679683 M * Bertl etteyafed: pong 1206679689 M * etteyafed Its late for you to be up still isn't it 1206679719 M * Bertl well, yeah, already went to bed, couldn't sleep ... 1206679735 M * etteyafed Any ideas what types of permissions my vserver might need to create TCP sockets? 1206679761 M * Bertl tcp sockets are allowed by default, nothing special required 1206679844 M * etteyafed It must be selinux then. This system is going to be a nightmare I think. I might end up just putting this on A standalone host. But I had really wanted the added security and flexibility of a vserver. 1206679873 M * Bertl what is the reason for the selinux stuff? 1206679898 M * etteyafed And this DB doesn't need all the power of the machine that I have to put it on. Well I need selinux for SE-PostgreSQL. 1206679935 M * etteyafed ANd if properly configured it would make the system (with vservers) very secure. 1206679963 M * Bertl well, then you should go through the 'configure properly' part :) 1206680024 M * etteyafed But setting it all up is not easy. SELinux can be a bit of a pain. Basically it is both a blessing and a curse to have SELinux be independent of the vserver. 1206680067 M * etteyafed If only it would just behave as it does in a regular host. 1206680122 M * Bertl well, actually it does, and that is probably what makes it harder for you 1206680137 M * Bertl i.e. there is no SElinux virtualization 1206680402 M * etteyafed Well that's what I meant. If only SELinux in the vserver acted like the vserver WAS the host then that would solve my problems, but instead I have to manually set all kinds of labels and edit policy by hand, and define custom transitions, etc. I could always use virtualization, but that would partially defeat the purpose of using the vserver, or not using a whole system for the job. 1206680445 M * Bertl OTOH, if it was completely virtualized, you wouldn't be able to have host policies for the guests 1206680458 M * Bertl i.e. each guest would make its own policy 1206680734 M * etteyafed Well yeah, and that might be less of a headache than figuring out how to do this in a way that afik has not been done before. I just keep thinking that it will be worth it once its done an I can have 2-3 DBs and their backups running on the same HW and be able to run them from one secure host that doesn't even need to be network accessible. 1206680804 M * Bertl I think it was done several times before, but not well documented, but except for the issues the private namespaces (and chroot) will cause for SElinux, it should be straight forward 1206680881 M * etteyafed Well I wish I could just sit down for a cup of coffee (or even tea) with someone that has done this before and go over the issues they encountered and how they were solved. 1206680906 M * Bertl checking the IRC logs might help there (or even browsing the mailing list) 1206681001 M * etteyafed That's what I have been up to actually. I have even considered looking for a consultant that has the knowledge on this subject we need, but I don't think its in the budget. Its my job anyway. 1206681407 M * Bertl k, I think I'll head back to bed now ... good luck! 1206681416 N * Bertl Bertl_zZ 1206681459 M * etteyafed nite 1206685049 J * cryptronic ~oli@p54A3B79C.dip0.t-ipconnect.de 1206686349 J * Slydder ~chuck@194.59.17.53 1206687211 J * hparker ~hparker@208.188.248.81 1206687259 J * ntrs ~ntrs@77.29.64.86 1206687896 Q * ntrs Ping timeout: 480 seconds 1206688256 Q * Linus Quit: I'll be back 1206688304 J * sharkjaw ~gab@64.28.12.166 1206688424 Q * cryptronic Quit: Leaving. 1206688512 J * JonB ~NoSuchUse@77.75.164.169 1206689577 J * ntrs ~ntrs@77.29.67.185 1206689869 Q * hparker Quit: Read error: 104 (Peer reset by connection) 1206689896 J * hparker ~hparker@208.188.248.81 1206690045 Q * JonB Quit: This computer has gone to sleep 1206690296 J * dna ~dna@86-246-dsl.kielnet.net 1206690969 J * tobifix ~tobifix@IVV7KNALLER.UNI-MUENSTER.DE 1206690987 M * tobifix good morning folks ;) 1206691303 J * mrfree ~mrfree@host1-89-static.40-88-b.business.telecomitalia.it 1206692083 J * JonB ~NoSuchUse@130.227.63.19 1206692452 Q * pusling Remote host closed the connection 1206692459 J * pusling pusling@77.75.162.71 1206692465 Q * nebuchadnezzar Remote host closed the connection 1206692898 J * pusling_ pusling@77.75.162.71 1206692910 Q * nkukard Quit: Leaving 1206692973 Q * pusling Read error: Connection reset by peer 1206693229 J * gebura ~gebura@77.192.186.197 1206693434 J * nkukard ~nkukard@196.212.73.74 1206693460 Q * hparker Quit: Read error: 104 (Peer reset by connection) 1206693583 M * gebura morning ! 1206693772 N * pusling_ pusling 1206693857 Q * nkukard Remote host closed the connection 1206693890 J * nkukard ~nkukard@196.212.73.74 1206694071 Q * nkukard Remote host closed the connection 1206694099 J * nkukard ~nkukard@196.212.73.74 1206694766 Q * pusling Ping timeout: 480 seconds 1206695034 J * pusling pusling@77.75.162.71 1206695174 J * virtuoso_ ~s0t0na@ppp91-122-27-116.pppoe.avangarddsl.ru 1206695371 Q * pusling Read error: Connection reset by peer 1206695373 J * pusling_ pusling@77.75.162.71 1206695523 Q * virtuoso Ping timeout: 480 seconds 1206695919 J * ktwilight_ ~ktwilight@89.113-66-87.adsl-dyn.isp.belgacom.be 1206696074 N * pusling_ pusling 1206696173 Q * ktwilight Ping timeout: 480 seconds 1206696916 J * pusling_ pusling@77.75.162.71 1206696920 N * DoberMann[ZZZzzz] DoberMann 1206696951 Q * pusling Read error: Connection reset by peer 1206697184 J * ftx_ ~ftx@dslb-084-060-251-188.pools.arcor-ip.net 1206697256 J * Hiaslboy ~wieser@143.224.23.205 1206697458 N * pusling_ pusling 1206697713 J * Julius ~julius@p57B255FB.dip.t-dialin.net 1206697988 J * ktwilight ~ktwilight@186.115-66-87.adsl-dyn.isp.belgacom.be 1206698233 Q * ktwilight Read error: Connection reset by peer 1206698241 Q * pusling Read error: Connection reset by peer 1206698248 J * pusling pusling@77.75.162.71 1206698267 J * ktwilight ~ktwilight@99.94-66-87.adsl-dyn.isp.belgacom.be 1206698398 Q * ktwilight_ Ping timeout: 480 seconds 1206698575 Q * pusling Read error: Connection reset by peer 1206698578 J * pusling_ pusling@77.75.162.71 1206698680 N * pusling_ pusling 1206699977 Q * mrfree Remote host closed the connection 1206700550 J * ftx__ ~ftx@dslb-084-060-234-060.pools.arcor-ip.net 1206700863 J * Punkie ~Punkie@goc.coolhousing.net 1206700893 Q * ftx_ Ping timeout: 480 seconds 1206702621 Q * Hiaslboy Ping timeout: 480 seconds 1206703209 J * ktwilight_ ~ktwilight@87.66.205.242 1206703445 J * virtuoso ~s0t0na@ppp91-122-94-49.pppoe.avangarddsl.ru 1206703556 Q * ktwilight Ping timeout: 480 seconds 1206703853 Q * virtuoso_ Ping timeout: 480 seconds 1206704019 Q * Slydder Quit: Leaving. 1206704546 J * Slydder ~chuck@194.59.17.53 1206705347 J * lilalinux ~plasma@80.69.41.3 1206705766 Q * ntrs Ping timeout: 480 seconds 1206706134 J * mrfree ~mrfree@host1-89-static.40-88-b.business.telecomitalia.it 1206706432 Q * mrfree 1206708627 N * ensc Guest2911 1206708627 Q * Guest2911 Remote host closed the connection 1206708637 J * ensc ~irc-ensc@77.235.182.26 1206708880 M * daniel_hozac ensc: ping 1206709559 Q * lilalinux Remote host closed the connection 1206710691 J * yarihm ~yarihm@whitehead2.nine.ch 1206712687 J * virtuoso_ ~s0t0na@ppp78-37-176-25.pppoe.avangarddsl.ru 1206712691 Q * virtuoso Read error: Connection reset by peer 1206712884 Q * tobifix Quit: Leaving 1206715093 Q * Aiken Remote host closed the connection 1206715501 J * dowdle ~dowdle@67-42-224-198.blng.qwest.net 1206715508 Q * ||Cobra|| Remote host closed the connection 1206715550 N * Bertl_zZ Bertl_oO 1206715889 J * ||Cobra|| ~cob@pc-csa01.science.uva.nl 1206717303 Q * eyck_ Ping timeout: 480 seconds 1206717600 J * eyck 7KtCsukh@nat05.nowanet.pl 1206717648 M * glen_ how to build redhat 7.3 vserver guest? :) 1206717744 J * dreamind ~dreamind@C2107.campino.wh.tu-darmstadt.de 1206717781 N * dreamind Guest2919 1206717855 N * Guest2919 dreamind 1206717857 M * dreamind Hi 1206718130 Q * Punkie Quit: Odcházím 1206718285 M * glen_ or fc2 1206718290 M * glen_ # vserver fc2 build -n fc2 --context 72 -m yum -- -d fc2 1206718290 M * glen_ cp: cannot stat `/etc/vservers/.defaults/vdirbase/.pkg/fc2/yum/etc/yum.conf': No such file or directory 1206718294 M * glen_ # vserver fc2 build -n fc2 --context 72 -m apt-rpm -- -d fc2 1206718294 M * glen_ command '/usr/lib64/util-vserver/vapt-get-worker' (pid 32127) exited with signal 11; following it... 1206718298 M * glen_ /usr/lib64/util-vserver/functions: line 716: 32126 Segmentation fault "$@" 1206719118 M * glen_ vyum fails: 1206719118 M * glen_ ++ /usr/sbin/vyum /etc/vservers/fc2 -- -y install glibc libgcc basesystem glibc-common 1206719121 M * glen_ Error: Missing Dependency: libgcc is needed by package glibc 1206719124 M * glen_ Error: Missing Dependency: basesystem is needed by package glibc 1206719126 M * glen_ Error: Missing Dependency: glibc-common = 2.3.3-27.1 is needed by package glibc 1206719276 Q * gebura Quit: Quitte 1206719538 Q * virtuoso_ Ping timeout: 480 seconds 1206719611 Q * Slydder Quit: Leaving. 1206719672 J * julius_ ~julius@p57B26EF1.dip.t-dialin.net 1206720000 Q * sharkjaw Remote host closed the connection 1206720098 Q * Julius Ping timeout: 480 seconds 1206720239 Q * yarihm Quit: This computer has gone to sleep 1206720983 J * bronson ~bronson@adsl-68-122-117-135.dsl.pltn13.pacbell.net 1206721141 M * Bertl_oO glen_: you need a 'proper' repository in any case 1206721156 M * Bertl_oO glen_: but the segfault would point to broken tools 1206721198 N * Bertl_oO Bertl 1206721213 M * glen_ well, i fould fc2 mirror with yum, it fetched pkgs and stuff, but the dep error is there 1206721223 M * glen_ probably yum has something missing, or too new 1206721230 J * docelic ~docelic@78.134.200.134 1206721262 M * Bertl glen_: maybe the repository is out of sync? 1206721282 M * Bertl glen_: check if the required packages are on that repository 1206721859 J * virtuoso ~s0t0na@ppp78-37-250-134.pppoe.avangarddsl.ru 1206722420 M * glen_ they were... 1206722438 M * glen_ but nevermind, downloading iso and using virtualbox to scratch install 1206722470 M * Bertl yep, for such old distros (which are not really supported anymore) that is probably the best choice 1206722819 M * Bertl weasel: http://vserver.13thfloor.at/Experimental/patch-2.6.24.4-vs2.3.0.34.diff 1206722834 M * Bertl (passes testme.sh and testfs.sh for ext2/3) 1206723069 N * DoberMann DoberMann[PullA] 1206723333 M * glen_ eh. is it tought to start vserver from some disk image? :) 1206723351 M * Bertl well, you can do a loop mount of that image 1206723368 M * Bertl but it is advised to copy the data over 1206723410 M * glen_ it would be then more like vmware, vserver in single file 1206723429 M * glen_ perhaps i could use scripts/prestart or sth to mount stuff? 1206723432 M * Bertl well, except that it doesn't work this way, but yeah, you can do that 1206723457 M * Bertl just put the proper loop mount in the guest config's fstab 1206723500 M * glen_ well it's little complicated, it's hard disk image 1206723521 M * Bertl so it needs a little offseting, but that's it 1206723539 M * glen_ right 1206723624 Q * esa Quit: Coyote finally caught me 1206723724 M * glen_ how to figure out that offset thing? 1206723752 M * Bertl there is a script on my server which does that for qemu harddisk images 1206723773 M * Bertl http://vserver.13thfloor.at/Stuff/QEMU/ 1206723795 M * glen_ i know that multipath-tools had some tool to do so too, it created dm entries 1206723800 M * glen_ kpartx 1206723959 M * glen_ /home/glen/.VirtualBox/VDI/Fedora Core 2.vdi: innotek VirtualBox Disk Image 1206723960 M * glen_ eh 1206724164 Q * docelic Quit: http://www.spinlocksolutions.com/ 1206725650 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d 1206725746 Q * dreamind Quit: dreamind 1206727076 M * glen_ any ideas in what package fc2 has xxd (it's in some vim pkg) 1206727115 M * Bertl it's in vim-common here too (mandriva) 1206727653 Q * JonB Ping timeout: 480 seconds 1206727906 M * glen_ found disk start offset of type=fixed image: http://pld.pastebin.com/f382f611c 1206727917 J * yarihm ~yarihm@access.rz1.unixwerkz.net 1206727965 M * Bertl good, now calculate the offset of the partition, and add it to the loop mount 1206728110 M * glen_ combining your script now :) 1206728173 J * Piet ~piet@tor.noreply.org 1206728174 J * tobifix ~tobifix@muedsl-82-207-233-232.citykom.de 1206728188 M * tobifix hey folks 1206728197 M * Bertl evening tobifix! 1206728452 M * glen_ deem 1206728453 M * glen_ # losetup -o 64512000 -s 512483328 2 /dev/loop1 1206728453 M * glen_ losetup: invalid option -- s 1206728462 M * glen_ no size support 1206728477 M * Bertl ah, well, you can do without 1206728488 M * glen_ but only if i use only onec partition i guess 1206728509 M * glen_ ah. no. if fs is not corrupted it will not access beyond fs blocks 1206728515 M * Bertl the filesystem does usually not go beyond the partition 1206728529 M * Bertl but of course, an losetup which handles the size is better 1206728548 M * glen_ eh, i can't losetup losetup dev, right? 1206728553 M * glen_ # losetup -o 64512000 /dev/loop0 1206728553 M * glen_ /dev/loop0: [fe06]:38341405 (Fedora Core 2.vdi), offset 8704 1206728557 M * glen_ it displayed me info instead :) 1206728559 M * Bertl sure you can 1206728590 M * Bertl that is what my scrip does ... again I have a 'better' losetup :) 1206728604 M * glen_ ok, mine will display status as you can see 1206728626 M * glen_ however i don't need losetup losetup, i just can +offset :) 1206728698 M * glen_ eh. some evil idea. could make /sbin/mount.vdi and use fstype=vdi in fstab :D 1206728940 M * glen_ 20:28:53 root[pts/2]@wintersunset /home/qemu# ./vdioffset.sh Fedora\ Core\ 2.vdi 1206728943 M * glen_ Fixed disk: Fedora Core 2.vdi 1206728946 M * glen_ losetup -o 40960 Fedora Core 2.vdi 1206728948 M * glen_ losetup -o 64520704 Fedora Core 2.vdi 1206728951 M * glen_ losetup -o 577004032 Fedora Core 2.vdi 1206728965 M * glen_ script from here: http://glen.alkohol.ee/pld/vdioffset.sh 1206729135 M * glen_ any suggestion on the usage? 1206729150 M * glen_ mount -t vdi /path/to/image -o partition=1 ? 1206729175 M * glen_ part=n Select partition number n from the device. Only makes sense for CDROMS. Defaults to not parsing the partition table at all. 1206729192 M * glen_ man mount ... -t hfs uses part=n, will do the same 1206729206 M * glen_ but udf has partition= ... 1206729357 N * DoberMann[PullA] DoberMann 1206729533 M * tobifix Bertl, i am doing the setup right now ;) 1206729540 M * tobifix thought about something like this 1206729549 M * tobifix 1. partition, /boot with 300m 1206729559 M * tobifix 2. partition, swap, 2gb 1206729570 M * tobifix 3. partition, / 10gigs 1206729576 M * tobifix the rest is yours 1206729578 M * tobifix ^^ 1206729716 J * larsivi ~larsivi@144.84-48-50.nextgentel.com 1206729763 Q * derjohn Server closed connection 1206729774 J * derjohn ~derjohn@80.69.41.3 1206729984 M * tobifix just need to tell me the mountpoint for the rest space 1206730048 Q * balbir Ping timeout: 480 seconds 1206730698 Q * yarihm Quit: This computer has gone to sleep 1206730703 J * balbir ~balbir@122.167.178.19 1206731805 M * glen_ yay. worx 1206731828 M * glen_ 21:16:55 root[pts/2]@wintersunset /home/qemu# mount -t vdi ./Fedora\ Core\ 2.vdi -o partition=3 /mnt/tmp/ -r 1206731831 M * glen_ 21:17:00 root[pts/2]@wintersunset /home/qemu# df /mnt/tmp/ 1206731833 M * glen_ Filesystem Type Size Used Avail Use% Mounted on 1206731836 M * glen_ /home/qemu/Fedora Core 2.vdi ext3 1.5G 587M 813M 42% /mnt/tmp 1206731966 J * DanielW ~danielw@f050064231.adsl.alicedsl.de 1206731968 M * DanielW hi 1206732041 M * DanielW were does booting on the guest begin? is /sbin/init executed or does it begin at /etc/init.d/rc ? 1206732068 M * arachnist what's -t vdi? 1206732070 M * glen_ Bertl, if want, then commited to cvs the helper: http://cvs.pld-linux.org/cgi-bin/cvsweb.cgi/SOURCES/mount.vdi 1206732080 M * glen_ arachnist, virtualbox disk image file 1206732083 M * arachnist oh 1206732101 M * glen_ just had suchj evil idea to mount /vserver/fc2 from image file 1206732135 M * glen_ let's hope it was worth of hacking and adding fstab entry for / will work 1206732637 M * glen_ /bin/mount -n -t ext3 -o rw,nodev,offset=577004032 /vservers/fc2.vdi . 1206732641 M * glen_ where the nodev came from? 1206732649 M * glen_ 21:30:44 root[pts/2]@wintersunset vservers/fc2# head fstab 1206732649 M * glen_ /vservers/fc2.vdi / vdi fstype=ext3,partition=3 0 0 1206732724 M * glen_ as it creates: /etc/rc.d/rc: line 87: /dev/null: Permission denied 1206732833 J * yarihm ~yarihm@whitehead2.nine.ch 1206732953 M * glen_ um. i can't enter that vserver 1206732959 M * glen_ # vserver fc2 enter 1206732959 M * glen_ vlogin: openpty(): No such file or directory 1206732960 M * glen_ 1206733064 J * JonB ~NoSuchUse@77.75.164.169 1206733332 M * glen_ mounted with -o dev 1206734361 M * glen_ hehe 1206734361 M * glen_ Mar 28 21:59:05 wintersunset kernel: grsec: From 192.168.2.3: denied resource overstep by requesting 92961307398144 for RLIMIT_STACK against limit 8388608 for /usr/lib64/util-vserver/secure-mount[secure-mount:19425] uid/euid:0/0 gid/egid:0/0, parent /usr/sbin/vserver[vserver:19365] uid/euid:0/0 gid/egid:0/0 1206734367 M * glen_ Mar 28 21:59:05 wintersunset kernel: grsec: From 192.168.2.3: denied resource overstep by requesting 92961307398144 for RLIMIT_STACK against limit 8388608 for /usr/lib64/util-vserver/secure-mount[secure-mount:19425] uid/euid:0/0 gid/egid:0/0, parent /usr/sbin/vserver[vserver:19365] uid/euid:0/0 gid/egid:0/0 1206734443 Q * larsivi Ping timeout: 480 seconds 1206734444 M * glen_ Bertl, otherwise it seems working :) 1206734598 M * glen_ how about using dhcp inside vserver? possible? 1206735138 M * glen_ blah. losetup dev doesnt' get released :( 1206735206 Q * yarihm Quit: This computer has gone to sleep 1206736007 N * DoberMann DoberMann[PullA] 1206738333 Q * DanielW Read error: Connection reset by peer 1206738658 J * puck ~puck@leibniz.catalyst.net.nz 1206738703 M * puck Hey, I'm running 2.3.0.34 and I've tried sticking an IPv6 address in my interfaces/1/ip file, and the dummy0 interface in interfaces/1/dev, but I get this error: 1206738712 M * puck Broadcast can be set only for IPv4 addresses 1206738712 M * puck naddress: vc_net_add(): Invalid argument 1206738719 M * puck Any suggestions for what I'm doing wrong? 1206738746 M * puck This is Debian Etch, so util-vserver 0.30.212-1. 1206738758 M * Bertl glen_: dhcp (client) is below ip protocol, so no, not inside a guest, but you can grab an ip for your guest on the host (with dhcp) 1206738796 M * Bertl puck: ancient utils, upgrade to 0.30.214+ and you should be fine 1206738819 M * puck Ahh, the page on the wiki needs to be updated then, it says 212 is fine. :( 1206738830 M * Bertl for ipv6? are you sure? 1206738850 M * puck http://linux-vserver.org/IPv6 1206738884 M * puck So the whole page is out of date, but... 1206738919 M * Bertl well, double check it (maybe with daniel_hozac too) and please update the page accordingly 1206738932 M * puck I'm upgrading now. 1206738934 M * Bertl it might be that 0.30.212 was fine with bruno's patches 1206738972 M * puck Quite possibly, definitely now that I'm using 0.30.214, the vserver started and has an IPv6 address! 1206738975 M * puck Thank you! 1206738989 M * puck It even works. :) 1206738992 M * Bertl excellent! please update/extend the wiki page 1206739005 M * puck Sure 1206739008 A * puck creates and account 1206739031 M * tobifix Bertl, got my msg concerning the partitions? 1206739092 M * Bertl yep, looks good 1206739125 M * Bertl again, root/swap should be on raid 1 (at least the system partition) 1206739195 M * puck done 1206739271 P * dowdle Konversation terminated! 1206739328 M * glen_ Bertl, there's some problem with umounting and releasing loopback 1206739349 M * glen_ if i mount from host, and umount, loopdev is released, but not with vserver root being loopback dev 1206739388 M * Bertl if you put it into the guest config's fstab (as I suggested the first time) it should get unmounted automatically 1206739413 M * glen_ it's unmounted but loopdev is not released 1206739421 M * glen_ test on host that loopdev is released: http://pld.pastebin.com/f3ab0eed 1206739494 M * glen_ vserver test: http://pld.pastebin.com/m605771ee 1206739519 M * glen_ 23:25:05 root[pts/2]@wintersunset vservers/fc2# cat /etc/vservers/fc2/fstab 1206739519 M * glen_ /vservers/fc2.vdi / vdi fstype=ext3,partition=3 0 0 1206739524 M * Bertl hmm, did you try with util-vserver 0.30.215 yet? 1206739539 M * glen_ nop sorry, too old llh to compile it 1206739568 M * Bertl IIRC, daniel_hozac fixed/added something there (in any case, check the changelogs) 1206739594 M * Bertl but you can work around this by assigning and releasing the loop device yourself in the pre/post scripts 1206739614 M * Bertl and putting just a normal mount with fixed loop device in the fstab 1206739617 M * glen_ it's hard to figure out which dev is used 1206739645 M * glen_ i'll give .215 a try 1206739658 M * Bertl well, you can 'define' which is used if you assign it 1206739688 M * Bertl e.g. designate loop5 for you guest, do the losetup in pre/post scripts, and use /dev/loop5 for the mount 1206739785 M * glen_ not that critical especially if .215 fixes it 1206739817 M * Bertl don't know for sure, but I have something in the back of my head ... 1206739864 M * glen_ already building 1206740220 J * mrfree ~mrfree@host44-50-dynamic.16-79-r.retail.telecomitalia.it 1206740511 M * Supaplex waldi: mind if I /msg you? (debian, spi, vserver, funding etc) 1206740548 M * glen_ Bertl, .215 same problem 1206740565 M * glen_ Bertl, but isn't it's umount who does losetop -d? 1206740594 M * glen_ Bertl, sorry. oops. still tested with .214 1206740595 M * waldi Supaplex: write an email, i'm away now 1206740625 M * Supaplex waldi: I'll cc: what I sent neilm@ spi 1206741159 M * glen_ Bertl, k. this time realy tested with .215, losetup -d is not done, i guess this is because mount -n is used and the "loop" information is lost and mount doesn't know that it should do losetup -d internally 1206741233 M * glen_ on host the /proc/mounts contains: /dev/loop1 /mnt/tmp ext3 rw,data=ordered 0 0 1206741251 M * Bertl could be ... have to think about this, it might need a little help from the kernel side 1206741290 M * glen_ 72: /dev/loop0 /vservers/fc2 ext3 rw,nodev,data=ordered 0 0 1206741317 M * glen_ that's vnamespace 72 ccat /proc/mounts (72 is vserver xid) 1206741334 M * glen_ so it's not lost at this point. i think 1206741362 M * glen_ ah. mtab should be compared, as mount prefers that info 1206741397 M * glen_ i can check only host mtab, don't know what it sees at vserver shutdown 1206741406 M * glen_ probably only /proc/mounts? 1206741427 M * glen_ on host its: 1206741427 M * glen_ /home/qemu/Fedora\040Core\0402.vdi /mnt/tmp ext3 rw,loop=/dev/loop1 0 0 1206741444 M * Bertl it should contain the same for the guest 1206741461 M * Bertl (the /proc/mounts) 1206741462 M * glen_ but the mtab for guest is fake, and afaik not used by util-vserver scripts 1206741485 M * Bertl but, IIRC, daniel_hozac was adding loop support at some point (for the root fs) 1206741504 M * Bertl (it uses a special mount command) 1206741515 Q * dna Quit: Verlassend 1206741520 M * glen_ yep. if i edit mtab and remove 'loop' from options line, then umount is not doing losetup -d 1206742022 J * yarihm ~yarihm@84-75-103-252.dclient.hispeed.ch 1206742248 Q * bonbons Quit: Leaving 1206742389 M * daniel_hozac umount is not called on vserver stop unless you disable the namespace. 1206742409 Q * JonB Quit: Leaving 1206742466 M * glen_ um? if i stop vserver the xid is gone, ie the dir is gone from /proc/virtual 1206742481 M * glen_ how to see if i disabled or not the namespace? 1206742502 M * daniel_hozac ls -l /etc/vservers/{guest,.defaults}/nonamespace 1206742535 M * glen_ ls: cannot access /etc/vservers/fc2/nonamespace: No such file or directory 1206742535 M * glen_ ls: cannot access /etc/vservers/.defaults/nonamespace: No such file or directory 1206742570 M * glen_ so i should echo 1 > /etc/vservers/fc2/nonamespace ? (or just vice versa)? ... confused 1206742580 M * daniel_hozac i doubt that would fix it. 1206742672 M * glen_ you are right. it didn't 1206742708 M * glen_ imho the problem is simple: for umount(1) to losetup -d, it needs to find 'loop' from options line, /etc/mtab has it, /proc/mounts doesn't. and for vserver mounts there's no mtab 1206742733 M * daniel_hozac it's not at all that simple. 1206742760 M * daniel_hozac umount is not called normally. 1206742810 M * daniel_hozac in order to properly support -o loop, you'd need to scan the guest's fstab on stop, find the associated /dev/loopX, unmount the filesystems that use them, losetup -d the devices. 1206742824 M * daniel_hozac it's a _lot_ of code for quite a special case, IMHO. 1206742861 M * Bertl agreed, but it can be done (by glen_) in a shutdown script, no? 1206742862 M * glen_ but the problem itself is simple. ie the explanation why umount does it automatically for host, and not for guests (where mtab is fake) 1206742880 M * Bertl glen_: no, not at all 1206742886 M * daniel_hozac umount is not called. 1206742891 M * daniel_hozac so it has _nothing_ to do with it. 1206742895 M * Bertl glen_: in the typical case, unmount is not even used for guests 1206742923 M * glen_ ok vmount then 1206742939 M * daniel_hozac again, not relevant. 1206742946 M * Bertl glen_: nope, the mount vanishes with the namespace 1206742946 M * daniel_hozac mounts are destroyed when the namespace is destroyed. 1206742971 M * glen_ i see. that explains a lot. otherwise i was confused how come no umount(1) 1206742976 M * Bertl okay, off to bed now ... have a good one everyone! 1206742979 M * daniel_hozac good night Bertl! 1206742986 N * Bertl Bertl_zZ 1206743002 M * glen_ daniel_hozac, for that shutdown loop option scan there needs to made changes for the loop to appear in guest mtab, as currently it's not there either 1206743069 M * glen_ correct? 1206743072 M * daniel_hozac you can't trust the mtab. it's in the guest. 1206743080 M * daniel_hozac you ought to parse the fstab. 1206743268 M * glen_ i can't add "loop" to fstab, then my mount.vdi is not invoked 1206743298 M * glen_ if losetup had also xid context, that would make losetop also dissapear on xid destroy maybe 1206743656 Q * xdr_ Read error: Connection reset by peer 1206743988 J * hparker ~hparker@208.188.248.81 1206744050 Q * mrfree Quit: Leaving 1206744239 J * Aiken ~james@ppp121-45-192-61.lns1.bne1.internode.on.net 1206744457 J * xdr ~xdr@199-173-96-87.cust.blixtvik.se 1206744796 J * mattzerah ~matt@pool2-134.dyn.winshop.com.au 1206744968 Q * mattzerah 1206745886 Q * maddoc Remote host closed the connection 1206745888 J * maddoc maddoc@social.ostruktur.com 1206747583 Q * tobifix Remote host closed the connection 1206748478 J * fuqz damn@secure.signalstorm.org