1170892828 M * daniel_hozac that's unlikely to start any service though... 1170892867 M * daniel_hozac if you just want a shell inside, vserver ... start --rescue /bin/bash should do it. 1170892868 M * litage daniel_hozac: it'll tell me if the problem exists in an "extra" service (ie: one that isn't required to boot) 1170892886 M * daniel_hozac there are no boot services when using the plain initstyle... 1170892899 M * daniel_hozac and if anything, the problem would be with the last script started. 1170892936 M * daniel_hozac s/plain/sysv/ 1170893223 M * litage disabling services one at a time, starting from S99... 1170893250 M * litage daniel_hozac: it seems webmin is the problem 1170893257 M * daniel_hozac note that appendind exit 0 would make it succeed in any case... 1170893258 M * litage ...webmin 1170893262 M * daniel_hozac +english 1170893283 M * litage +english? 1170893293 M * litage +engrish! 1170893517 M * litage daniel_hozac: why does starting up a vserver fail if a single init script fails? 1170893539 M * daniel_hozac not a single, the last one. 1170893550 M * daniel_hozac and that's only because of the way the sarge initscript is written. 1170893566 M * daniel_hozac it returns failure. 1170893581 M * litage ah 1170893682 J * ensc ~irc-ensc@p54B4E55B.dip.t-dialin.net 1170893839 M * litage daniel_hozac: thanks for helping me out with this, i really appreciate it 1170893846 M * daniel_hozac you're welcome! 1170893865 Q * tso Quit: My damn controlling terminal disappeared! 1170893872 M * litage daniel_hozac: all's fixed now. it seems the problem is that webmin was trying to bind to port 80, but apache already had it 1170893879 M * litage :) 1170894166 M * daniel_hozac good 1170894331 J * Aiken ~james@ppp220-70.lns2.bne1.internode.on.net 1170894342 M * Bertl morning Aiken! 1170894377 M * Aiken good morning 1170895301 Q * FireEgl Quit: ... 1170895896 M * baldy Bertl: can u code php? 1170895927 M * Bertl yup 1170896042 Q * meandtheshel1 Quit: Leaving. 1170896327 Q * Radiance Ping timeout: 480 seconds 1170896559 J * Radiance d28304fd24@halt.1984world.eu 1170896725 Q * Piet_ Quit: Piet_ 1170897100 M * Bertl okay folks ... off to bed now ... have a good one! 1170897108 M * daniel_hozac good night! 1170897134 M * Bertl have a good one too! 1170897139 N * Bertl Bertl_zZ 1170897208 J * donuto ~donuto@83.149.112.45 1170897216 M * donuto hello :) 1170897248 P * litage Leaving 1170897281 M * donuto i had small accident, i give cap_net_admin to vserver to make routing working, but when i did vserver test stop the whole machine lost network connection ! 1170897319 M * daniel_hozac as expected, if you didn't clean up your guests initscripts. 1170897346 M * donuto ahh 1170897371 M * donuto i must edit/remove network down scripts from the vserver? 1170897393 M * daniel_hozac well, you shouldn't give your guest CAP_NET_ADMIN in the first place. 1170897401 M * daniel_hozac that lets it do whatever it wants to the network stack. 1170897416 M * donuto i am only user of that vserver, no other person there 1170897465 M * donuto i had no option, else openvpn was not working with route setup 1170897480 M * daniel_hozac you can configure the route on the host. 1170897521 M * donuto openvpn also need tx and mtu setting, this cannot be stopped 1170897535 M * daniel_hozac i seriously doubt that. 1170897538 M * donuto else openvpn fails with exit code 255 1170897564 M * daniel_hozac but anyway, i need to get some sleep. good luck, have fun, and good night! 1170897579 M * donuto thanks, wish you a good sleep friend ! :) 1170897664 Q * rgl Quit: Fui embora 1170899026 Q * dmax Ping timeout: 480 seconds 1170899594 Q * s0undt3ch Ping timeout: 480 seconds 1170899736 J * s0undt3ch ~s0undt3ch@80.69.34.154 1170900285 Q * donuto Quit: leaving 1170900332 Q * Radiance Remote host closed the connection 1170903220 J * almak ~almak@willers.employees.org 1170903238 M * almak 1170903551 M * almak hi there 1170903597 M * almak quit 1170903601 Q * almak Quit: almak 1170904374 M * ntrs Is there any easy way to implement COW on an existing running host with multiple guests? 1170904994 Q * SNy Read error: Operation timed out 1170905354 T * * http://linux-vserver.org/ | latest stable 2.0.2.1, 2.0.3-rc1, 2.2.0-rc12, devel 2.1.1.7.1, 2.3.0.8, stable+grsec 2.0.2.1, 2.2.0-rc10 | util-vserver-0.30.212 | 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 ;) 1170905354 T * Bertl - 1170905592 M * cehteh ntrs: just build a COW enabled kernel and hashify the guest i think 1170907500 J * SNy dbf5a791ed@bmx-chemnitz.de 1170909915 Q * ntrs Remote host closed the connection 1170910665 N * DoberMann[ZZZzzz] DoberMann 1170911297 Q * Aiken Quit: Leaving 1170911394 J * ntrs ~ntrs@68-188-55-120.dhcp.stls.mo.charter.com 1170911547 J * Aiken ~james@ppp220-70.lns2.bne1.internode.on.net 1170912308 J * DoberMann_ ~james@90.38.158.36 1170912370 Q * DoberMann Read error: Operation timed out 1170913376 N * DoberMann_ DoberMann[PullA] 1170913528 N * Bertl_zZ Bertl 1170913532 M * Bertl morning folks! 1170913779 M * micah evening Bertl! 1170914027 M * Bertl hey micah! how's going? 1170914391 M * micah Bertl: quite tired and way too busy, but ok! how have you been? 1170914510 M * Bertl very busy recently too ... 1170914533 M * Bertl otherwise fine .. 2.2.0 should be done any day now ... 1170914583 M * Bertl btw, we had quite a number of complaints regarding the 2.6.18-3 kernel in debian .. you might want to emphasize the importance of fixing that one up RSN :) 1170914688 M * micah Bertl: yeah, I think thats finally moving now, I had problems with it too! 1170914731 Q * SNy Read error: Connection reset by peer 1170914984 M * Bertl micah: that's always a good start (when maintainers have issues with related stuff :) 1170915211 M * Bertl micah: on a different note, is there some progress in reintegrating the debian guest build 'addons' into util-vserver? 1170915215 J * SNy aab30ae996@bmx-chemnitz.de 1170915452 M * mEDI_S hi bertl 1170915506 M * mEDI_S is the 64bit stats problem fixed in the newst patch version? 1170915536 M * mEDI_S from 2.2er tree ;) 1170915735 M * Bertl sure :) 1170915753 M * mEDI_S fixed? 1170915756 M * Bertl yes 1170915767 M * Bertl of course, we fix such issues very quickly :) 1170915777 M * mEDI_S jeeeeaaaa a big thanx ;) 1170915782 M * Bertl you're welcome! 1170915801 M * Bertl could need some testing though, so go ahead and test drive :) 1170915804 M * mEDI_S you're welcome! heh 1170915838 M * mEDI_S na i muss see i build first a test kernel on my destop for 2-3 days 1170915889 M * Bertl well, just report back how it works :) 1170915992 M * cehteh is a patch for 2.6.20 on the way? 1170916018 M * cehteh just asking because i have to reboot and a kernel upgrade wont be bad ... i could use .19 too 1170916027 M * mEDI_S kk do i ;) 1170916123 M * Bertl cehteh: nope, it's already out :) 1170916152 M * Bertl cehteh: but note, the 2.6.20 is not much tested yet, so it might break quite easily 1170916156 M * cehteh heh i looked 2 hours ago or so 1170916160 Q * hallyn Read error: Operation timed out 1170916170 M * cehteh not a problem .. just my desktop machine 1170916175 M * Bertl cehteh: you didn't look in the right place then ... 1170916182 M * cehteh as long it wont shred my filesystems its all ok 1170916192 M * cehteh i looked at the wiki 1170916201 M * Bertl http://vserver.13thfloor.at/Experimental/ 1170916209 M * cehteh oki 1170916211 M * Bertl it' 1170916220 M * Bertl +s there since more than a day now 1170916242 M * Bertl but I guess there will be a pre2 shortly 1170916255 M * Bertl actually I could do one right now :) 1170916296 M * cehteh eh .. i am thought about 2.2 .. or is 2.3 at least stable enough that when it works i dont need to upgrade the kernel for the next half year? 1170916324 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.20-vs2.2.0-pre1.diff 1170916355 M * cehteh that directory is a mess 1170916368 M * Bertl no, it is (by default) ordered by time/date 1170916385 M * Bertl which really makes sense, if you consider that fixes and releases are stored there :) 1170916401 M * cehteh heh ok 1170916407 M * Bertl but you can change the sorting via your browser :) 1170916424 M * cehteh yes seen 1170916560 M * cehteh shall i wait for pre2 (important fixes?) 1170916575 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.20-vs2.2.0-pre2.diff 1170916586 J * hallyn ~xa@cpe-72-179-43-119.austin.res.rr.com 1170916590 M * cehteh doh .. you are fast ;) 1170916616 M * Bertl I know .. I try to level that with adding more and more workload :) 1170916630 M * cehteh heh 1170916695 M * cehteh cp /boot/config-2.6.14.5-vs2.0.1 .config ... i guess its really time for a kernel update ;) 1170917019 M * Bertl definitely .. you will have to select quite a few new options 1170917035 M * Bertl IIRC, even IDE was moved around between those versions 1170918130 Q * ard Ping timeout: 480 seconds 1170918412 M * cehteh yeah ... compiling ... 1170918434 M * cehteh 20 mins just for make oldconfig .. doh 1170918449 Q * Aiken Quit: Leaving 1170919394 M * cehteh linux-image-2.6.20vs2.2.0-vs2.2.0-pre2ct_2.6.20vs2.2.0-vs2.2.0-pre2ct-10.00.Custom_amd64.deb 1170919414 M * cehteh wtf ... debian make-kpkg autogenerated package name 1170919495 M * Bertl nice name, let me check if the patch is to blame (at least partially :) 1170919522 M * mEDI_S lol 1170919527 A * cehteh rebuild the ./debian dir 1170919532 M * Bertl yep, we have that wrong, happens everytime a new branch is started 1170919548 M * cehteh hehe 1170919551 M * Bertl it says: vs2.2.0-vs2.2.0-pre2 instead of vs2.2.0-pre2 1170919563 M * Bertl will fix that in place in a second 1170919566 M * cehteh i just install it ;) gives it some unique touch 1170919672 M * cehteh now it gets weirder ... 1170919675 M * cehteh update-grub 1170919675 M * cehteh Searching for GRUB installation directory ... 1170919675 M * cehteh No GRUB directory found. 1170919739 M * cehteh # lilo 1170919739 M * cehteh -bash: lilo: command not found 1170919759 M * cehteh .. how did the box initially booted? 1170920065 M * Bertl magic? 1170920240 M * cehteh prolly ... had to reinstall grub completely ... 1170920245 M * cehteh so reboot lets see 1170920257 M * cehteh err ... do i have to update util-vserver too? 1170920287 M * cehteh util-vserver-0.30.209 1170920287 M * cehteh ... mhm yes 1170921035 M * Bertl wouldn't hurt 1170921050 M * cehteh .. fsck takes ages ... 1170921075 M * cehteh but so far it booted at least 1170922283 Q * virtuoso Ping timeout: 480 seconds 1170922310 J * virtuoso ~s0t0na@80.253.205.251 1170922378 Q * dlezcano Ping timeout: 480 seconds 1170922585 M * cehteh # uname -a 1170922585 M * cehteh Linux starbase 2.6.20vs2.2.0-vs2.2.0-pre2ct #2 Thu Feb 8 08:29:12 CET 2007 x86_64 GNU/Linux 1170922599 M * cehteh thats uniqe ;) 1170922809 M * Bertl yep 1170922831 M * cehteh vserver nebulae start 1170922832 M * cehteh /lib/util-vserver/vserver.start: line 133: /usr/bin/nice: No such file or directory 1170922834 M * cehteh hmpf 1170922838 M * cehteh nice is there 1170922848 J * Aiken ~james@ppp102-23.lns1.bne1.internode.on.net 1170922863 M * cehteh is there some debug option for vserver? 1170924071 M * Bertl yes, quite a lot of options are there ... 1170924087 M * cehteh chbind: kernel does not provide network virtualization 1170924093 M * Bertl i.e. --debug for the userspace tools, and sysctl debug support for the kernel side 1170924095 M * cehteh aha ... 1170924111 M * Bertl did you compile the tools yourself? 1170924163 M * cehteh yes 1170924173 M * Bertl which version? 1170924198 M * cehteh util-vserver-0.30.212 1170924212 M * Bertl try with one of the 0.30.213-rc* 1170924222 M * cehteh oki 1170924241 M * Bertl did you enable legacy networking or so in the kernel config? 1170924285 M * Bertl (not that this shouldn't work, but I assume your kernel interface is too new for the tools) 1170924308 M * cehteh i just ran make oldconfig and fixed it up ... 1170924319 M * cehteh legacy is on 1170924327 M * Bertl is _now_ on? 1170924356 M * cehteh for the kernel i build yes 1170924368 M * cehteh kernel api and network api ... both legacy turned on 1170924374 M * Bertl ah, well, you could leave it off if you update the tools :) 1170924407 J * ard ~ard@82-197-200-127.dsl.cambrium.nl 1170924408 M * cehteh well .. shouldnt break when it is on or 1170924424 M * Bertl nope, just adds legacy stuff ... 1170924564 M * cehteh mhm where are the bleeding edge tools? 1170924888 M * Bertl sec 1170924898 M * cehteh np 1170924964 M * Bertl ftp://ftp.linux-vserver.org/pub/people/dhozac/p/uv/experimental/util-vserver-0.30.213-rc1.tar.bz2 1170925177 M * cehteh same problem 1170925195 J * lilalinux ~plasma@80.69.41.2 1170925205 A * cehteh checks the --debug log 1170925291 M * Bertl try with testme.sh especially with the -d option 1170925313 T * * http://linux-vserver.org/ | latest stable 2.0.2.1, 2.0.3-rc1, 2.2.0-rc12, devel 2.1.1.7.1, 2.3.0.8, stable+grsec 2.0.2.1, 2.2.0-rc10 | util-vserver-0.30.212 | 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 ;) 1170925313 T * Bertl - 1170925315 J * SNy 342273a4bd@bmx-chemnitz.de 1170925321 J * mnemoc ~amery@kilo105.server4you.de 1170925322 J * virtuoso ~s0t0na@80.253.205.251 1170925324 J * Johnsie ~jdlewis@jdlewis.org 1170925326 J * Aiken ~james@ppp102-23.lns1.bne1.internode.on.net 1170925437 M * cehteh http://www.pipapo.org/vserver.log 1170925500 M * Bertl nice one, but let's check with testme.sh first ... 1170925508 M * cehteh where is the testme.sh? 1170925532 M * cehteh (not in the util vserver tarball) 1170925533 M * Bertl http://vserver.13thfloor.at/Stuff/SCRIPT/testme.sh-0.17 1170925574 M * cehteh chcontext is working. 1170925575 M * cehteh chbind: kernel does not provide network virtualization 1170925575 M * cehteh chbind: kernel does not provide network virtualization 1170925575 M * cehteh chbind failed! 1170925575 M * cehteh Linux 2.6.20vs2.2.0-vs2.2.0-pre2ct #2 Thu Feb 8 08:29:12 CET 2007 x86_64 1170925576 M * cehteh Ea 0.30.213 236/glibc (DSa) 1170925578 M * cehteh VCI: 0002:0200 236 03008777 (TbLgnvsPHiW) 1170925580 M * cehteh --- 1170925588 M * cehteh tsk ... i dont want network virtualization 1170925648 M * cehteh could --enable-release be the error? 1170925657 M * Bertl how did you configure the tools? 1170925668 M * cehteh ./configure --prefix='' --enable-release 1170925683 M * Bertl never tried that :) 1170925696 M * cehteh only changes CFLAGS 1170925706 M * cehteh -g .. no even -O changed 1170925747 M * Bertl the interfaces are fine, v13,net,v21 1170925752 M * cehteh yes 1170925768 M * cehteh syscall(2) invocation: alternative 1170925770 M * Bertl could you upload the full output of testme.sh -d to paste.linux-vserver.org please? 1170925813 M * cehteh moment ... uninstalled the tools 1170925814 J * dna ~naucki@60-218-dsl.kielnet.net 1170925920 J * ensc ~irc-ensc@p54B4E55B.dip.t-dialin.net 1170925936 M * cehteh http://www.pipapo.org/vserver.testme.log 1170925949 T * harry http://linux-vserver.org/ | latest stable 2.0.2.1, 2.0.3-rc1, 2.2.0-rc12, devel 2.1.1.7.1, 2.3.0.8, stable+grsec 2.0.2.1, 2.2.0-rc12 | util-vserver-0.30.212 | 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 ;) 1170925989 M * Bertl harry: great! 1170926095 M * cehteh chbind -- echo test 1170926095 M * cehteh chbind: kernel does not provide network virtualization 1170926101 M * cehteh really weird 1170926106 Q * cdrx Read error: Operation timed out 1170926132 M * Bertl I'm absolutely unsure about the status of chbind 1170926165 M * Bertl i.e. if that was updated to use the new interfaces and/or what is required to detect the API ... 1170926194 M * Bertl I assume it uses the older ABI, which is probably disabled in the kernel, the -d should tell 1170926246 M * cehteh huh since when is chbind a shell script .. hehe i looked ages ago on the source 1170926262 M * cehteh yes i straced it 1170926284 M * cehteh execve("/lib/util-vserver/chbind-compat", 1170926368 M * cehteh guess what .. i install the old tools now :P 1170926501 M * Bertl well, they won't work either (without the legacy stuff in the kernel enabled) 1170926534 M * cehteh well it is enabled 1170926537 M * cehteh chbind: vc_set_ipv4root(): Invalid argument 1170926553 M * cehteh .. maybe my config is fucked up .. where there some changes? 1170926588 M * Bertl well, dynamic contexts were finally thrown out 1170926601 M * Bertl (there is still an option to enable them though) 1170926610 M * cehteh # /sbin/chbind --silent --ip 10.20.20.20/16 -- echo test 1170926610 M * cehteh chbind: vc_set_ipv4root(): Invalid argument 1170926628 M * cehteh i added a static context .. did that already with my server 1170926630 M * Bertl try to add --nid 42 1170926647 M * cehteh invalid option `nid'. 1170926654 M * cehteh (the old tools now) 1170926674 M * cehteh util-vserver-0.30.209 1170926680 M * Bertl 2.2.0 with default config and 0.30.213 is supposed to work out of the box 1170926704 M * Bertl a failure to do so, is considered a bug ... but I guess we have to wait for daniel_hozac to figure the details 1170926712 M * cehteh is the uname parsed somewhere? 1170926739 M * Bertl no, it shouldn't be 1170926754 M * Bertl only the VCI version and the actual kernel interface 1170926803 M * cehteh mhm 1170926848 M * cehteh well not urgend ... i talk with daniel_hozac later .. this is luckily only my desktop and the vserver is just a 32bit distcc which is not really important 1170926865 M * Bertl okay ... 1170926914 M * cehteh i guess it worked somewhere else .. so prolly a problem on my side 1170927445 A * waldi just boots 2.2.0-pre1 on his sparc ... 1170927488 J * ema ~ema@lart.galliera.it 1170927505 J * DavidS ~david@chello062178045213.16.11.tuwien.teleweb.at 1170927528 M * waldi readdir(): Bad address 1170927528 M * waldi readdir(): Bad address 1170927528 M * waldi readdir(): Bad address 1170927528 M * waldi readdir(): Bad address 1170927528 M * waldi readdir(): Bad address 1170927529 Q * waldi Killed (FloodServ ((FloodServ) Warning, you have triggered a network protection. Stop flooding!)) 1170927545 M * Bertl nice :) 1170927577 J * waldi ~waldi@bblank.thinkmo.de 1170927582 M * Bertl wb waldi! :) 1170927587 M * waldi *gna* 1170927607 M * waldi ;) 1170927639 M * waldi SCSI device sdb: write cache: disabled, read cache: enabled, doesn't support DPO or FUA 1170927642 M * waldi hmm 1170927669 M * Bertl sounds very sparc-ish :) 1170927685 M * waldi yep 1170927966 M * waldi hmm, I get many Z processes 1170928032 M * Bertl that's not good, 2.6.20? 1170928063 M * waldi yep 1170928080 M * Bertl means we got the reaper stuff wrong again 1170928101 M * waldi yep, the PPID it 0 1170928103 M * Bertl had that with the previous .20 release too, but I thought I fixed it this time 1170928109 M * waldi s/it/is/ 1170928115 M * waldi this looks really wrong 1170928127 M * waldi root 4786 0 0 09:38 ? Ss 0:00 /usr/sbin/sshd 1170928127 M * waldi root 4804 0 0 09:38 ? Ss 0:00 /usr/sbin/cron 1170928127 M * waldi buildd 4970 0 0 09:39 ? Z 0:00 [buildd] 1170928127 M * waldi buildd 4972 0 0 09:39 ? Ss 0:00 /usr/bin/perl /usr/bin/buildd 1170928131 M * Bertl IIRC, I added some reaper debugging stuff ... 1170928161 M * harry patch seems to work nicely :) 1170928166 M * harry compiles clean :) 1170928215 M * waldi Bertl: processes which are forked in the vserver gets the correct parent 1170928267 M * Bertl so which ones get the wrong one? the host processes? 1170928282 M * Bertl ah, yes, makes sense actually ... 1170928329 M * waldi the processes which are started during vserver $bla start 1170928340 M * waldi all daemons 1170928358 M * Bertl hmm ... 1170928380 M * Bertl let me check what debug options we have in place 1170928391 M * Bertl I assume you compiled with debugging enabled, yes? 1170928396 M * waldi no 1170928409 M * Bertl re-compile time then? 1170928421 M * waldi yep 1170928658 M * Bertl debug flags, xid=5,6,7 switch=0,1 should give us some info 1170928957 J * Piet hiddenserv@tor.noreply.org 1170929084 J * dlezcano ~dlezcano@blueice1n1.uk.ibm.com 1170929575 M * Bertl welcome dlezcano! 1170929590 M * dlezcano Hi 1170929707 Q * thunder1 Remote host closed the connection 1170929721 Q * Aiken Quit: Leaving 1170929730 J * thunder1 ~thu@219.163.49.81 1170929960 M * waldi Bertl: which debug options? 1170930049 M * matti Bertl: :) 1170930221 M * Bertl xid=5,6,7 switch=0,1 1170930238 M * Bertl those are the bit numbers, you have to or/add them 1170930257 M * Bertl and use them either via boot time kernel commandline or sysctl 1170930315 M * waldi okay 1170930479 A * waldi .o0( -j64 ... ) 1170930618 M * Bertl hmm? 1170930655 M * waldi make -j64 1170930688 J * Aiken ~james@ppp102-23.lns1.bne1.internode.on.net 1170930979 M * waldi fuck, ENOSPC 1170931027 M * Bertl *tsts* 1170932152 M * waldi okay, lets boot it 1170932466 J * chand ~chand@m167.net81-64-156.noos.fr 1170933050 M * waldi rogue reaper: fffff807ffad8960[5008,#1] <> fffff80000c2bb60[1,#0] 1170933083 M * Bertl #1 is cool :) 1170933113 M * waldi why? 1170933125 M * Bertl that's the spectator context 1170933153 M * waldi yeah 1170933162 M * waldi but what is this rogue reaper? 1170933190 M * Bertl just means that the parent is in a different context 1170933196 M * waldi ah 1170933209 M * Bertl vxwprintk((p->xid != reaper->xid) && (reaper != &init_task), 1170933297 M * Bertl but the reaper != init_task makes me wonder 1170933314 J * meandtheshell ~markus@85.124.39.2 1170933324 M * Bertl which basically means that fffff80000c2bb60[1,#0] has pid=1 1170933334 M * Bertl but is _not_ the init process 1170933343 M * waldi weird 1170933507 M * waldi may this be the problem why all this processes, which should have init as parent have pid 0 instead? 1170933599 M * Bertl zep 1170933621 M * Bertl zep: emphasized yep or wrong keyboard layout :) 1170933646 M * Bertl you have the kernel tree at hand? 1170933650 M * waldi yes 1170933670 M * Bertl let's try to replace init_task with something else 1170933694 M * waldi hmm, is init_task arch specific? 1170933719 M * Bertl yes, but here are the relevant locations: 1170933737 M * Bertl kernel/exit.c 598 1170933748 M * waldi because init_task is defined as "swapper" aka pid 0 1170933752 M * Bertl kernel/pid.c 70 1170933760 M * waldi in include/linux/init_task.h 1170933765 M * Bertl kernel/vserver/context.c 95 1170933773 M * Bertl kernel/vserver/context.c 96 1170933777 M * Bertl kernel/vserver/context.c 821 1170933839 M * Bertl and we replace it with: 1170933853 M * Bertl init_pid_ns.child_reaper 1170933867 M * Bertl ah, forget the kernel/pid.c one 1170933872 M * waldi static inline struct task_struct *child_reaper(struct task_struct *tsk) 1170933873 M * waldi { return init_pid_ns.child_reaper; 1170933873 M * waldi } 1170933894 M * Bertl you can use that one too 1170933965 M * Bertl but IMHO that should end up being &init_task again 1170933977 M * Bertl but we'll see, maybe we need something else there 1170933978 M * matti What's up? 1170933999 M * Bertl zombies, lots of 'em ... 1170934221 M * matti OMG. 1170934228 M * matti Run, hide. 1170934230 M * matti ;] 1170934237 M * matti Bertl: Coffee? 1170934244 M * matti 0xc0ffee? 1170934379 M * Bertl yes, tx! 1170934386 M * Bertl waldi: ah, that should work ... 1170934415 M * Bertl I missed the following: (in init/main.c) where we have: init_pid_ns.child_reaper = current 1170934438 M * Bertl Hollow: Unknown MySQL server host 'zelos.croup.de' (2) (zelos.croup.de)" 1170934453 M * Bertl Hollow: on linux-vserver.org 1170934514 M * waldi Bertl: what? 1170934589 M * waldi ah, i see 1170934656 M * waldi *reboot* 1170934710 M * waldi bah, wrong silo conf ... 1170934778 Q * Aiken Quit: Leaving 1170935766 M * waldi Bertl: yep, looks much better 1170935917 M * Bertl okay, let me upload a fixed version for testing then 1170935943 A * matti hands a cup of coffee to Bertl 1170936178 M * Bertl waldi: okay, pre3 should fix that in a proper way ... 1170937649 M * daniel_hozac cehteh: you have the legacy version enabled. that will disable the new chbind, and if you don't build the tools with legacy support, there's nothing to fallback to 1170937681 M * Bertl hmm, sure about that? 1170937696 M * daniel_hozac yep. 1170937705 M * Bertl I mean, at least he said that he hadn't enabled any legacy 1170937721 M * Bertl and the defaults now should not enable legacy either, no? 1170937724 M * daniel_hozac 0300_8_777 1170937738 M * daniel_hozac nope, but he used oldconfig, no? 1170937759 M * Bertl good point ... I just wonder what he enabled then ... 1170937789 M * Bertl but putting this aside, what keeps the tools (when compiled without the legacy stuff) from using the non-legacy interface? 1170937807 M * daniel_hozac it checks the VCI version. 1170937814 M * Bertl which is the same? 1170937832 M * daniel_hozac if it's less than 0x20000, the new chbind implementation is used. 1170937846 M * daniel_hozac s/less/greater/ 1170937856 M * daniel_hozac what? 1170937863 M * Bertl VCI: 0002:0200 1170937873 M * daniel_hozac but that's parsing proc, no? 1170937887 M * daniel_hozac vc_get_vci(COMPAT) will still return 0x100FF. 1170937914 M * Bertl only with the deprecated show legacy version 1170937923 M * daniel_hozac exactly, which is enabled. 1170937952 M * Bertl hum, how did that get enabled ... 1170937955 M * daniel_hozac that's what will break 0.30.210+ if you don't compile the utils with legacy APIs. 1170937965 M * Bertl not even a make oldconfig should be able to do that 1170937973 M * daniel_hozac how so? 1170937987 M * Bertl because it depends on legacy and is off by default 1170938011 M * daniel_hozac but if legacy and legacy_version is enabled in the old config, it should work, no? 1170938024 M * Bertl probably 1170938046 M * Bertl and why does chcontext work then, but chbind not? 1170938062 M * daniel_hozac the migrate API doesn't depend on such a recent version. 1170938076 M * daniel_hozac it was introduced in 0x10025 or something like that, IIRC. 1170938119 M * Bertl somehow I'm still lost ... 1170938140 M * Bertl we have a kernel (granted with legacy enabled) which has the latest ABIs 1170938155 M * Bertl and we have tools, with the latest ABI enabled (only) 1170938170 M * Bertl still we get a message that the network virtualization is not there? 1170938183 M * daniel_hozac yeah, since the kernel says it's older than it is. 1170938192 M * Bertl sounds fishy to me ... but more important, what can we do against that? 1170938204 M * Bertl remove the legacy version option? 1170938210 M * daniel_hozac well, IMHO, this is the behaviour the legacy version is meant to induce. 1170938220 M * daniel_hozac _why_ people enable it is still a mystery to me though... 1170938302 M * Bertl hmm, no, IMHO the legacy version shown in the kernel is there to work around the bug enrico introduced 1170938321 M * Bertl which will make older versions of util-vserver fail with newer kernels 1170938328 M * daniel_hozac well, right. 1170938347 M * Bertl the purpose is not to trick newer tools into using the older ABI :) 1170938400 M * daniel_hozac i'm not sure why anyone would still be using the ancient utils... 1170938438 M * Bertl well, might be because of existing scripts/wrappers or so 1170938460 M * Bertl which use older interfaces or whatever ... not that I would encourage that 1170938999 J * Radiance 775158846b@halt.1984world.eu 1170939015 M * Bertl welcome Radiance! 1170939087 M * Radiance hi Bertl :) 1170939547 M * Hollow Bertl: well, works here, but i have added the host to /etc/hosts for helios now .. 1170939648 M * Bertl okay, tx 1170939698 M * Bertl nap attack! .. back later ... 1170939707 N * Bertl Bertl_zZ 1170939800 J * cdrx ~legoater@blueice2n1.uk.ibm.com 1170941237 M * Radiance napnes :) 1170941627 M * Hollow Bertl_zZ: can we get a iattr_setat()? :) 1170942237 M * daniel_hozac why? 1170942832 M * Hollow i wouldnÄt have to chdir to vdirfd on every file during linking/copying 1170942861 M * daniel_hozac hmm? vc_set_iattr takes absolute paths, no? 1170942893 M * Hollow iirc relative ones work too 1170942903 M * daniel_hozac sure. 1170942948 M * Hollow well, i have to concatenate the vdir and the relative path nftw gives me to set the approriate flags, cwd is not vdir, but the template dir 1170942962 M * daniel_hozac right. 1170942980 M * Hollow lame. 1170942981 M * Hollow :) 1170942985 M * daniel_hozac can't you set the flags on the template file though? 1170942995 M * Hollow on link yes, on copy no 1170943006 M * daniel_hozac why are you setting the flags on copy? 1170943013 M * Hollow good question 1170943022 M * Hollow i even don't do it 1170943030 M * daniel_hozac so... :) 1170943043 M * Hollow yeah .. thanks for the hint :) 1170943076 M * daniel_hozac if you do need a vc_set_iattrat, it shouldn't be a problem AFAICT. 1170943151 M * Hollow yeah, but it should work now .. 1170943362 J * olivierk ~olivier@olivierk.org 1170943660 M * pflanze What do I have to do to see all network information? In older releases it has been enough to chcontext --ctx 1. 1170943686 M * pflanze "all" meaning the network connections of all vserver clients. 1170943911 M * daniel_hozac ncontext --nid 1 --migrate -- ... 1170944178 M * pflanze Thanks. 1170944198 M * pflanze What does --migrate mean? (it's not explained in --help) 1170944215 M * pflanze ah, migrate or create 1170944243 M * pflanze hm, doesn't work for me. 1170944269 M * pflanze chcontext --ctx 1 ncontext --nid 1 --migrate lsof -i => outputs nothing 1170944282 M * daniel_hozac kernel? 1170944283 M * pflanze same thing w/o chcontext 1170944288 M * pflanze 2.6.19.2-vs2.2.0-rc8.7 1170944367 M * daniel_hozac i'm getting similar results here on my testing tree. 1170944432 M * pflanze ncontext --nid 1005 --migrate chcontext --ctx 1005 lsof -i *does* work (with a vserver running in 1005) 1170944443 M * pflanze but with 1, nothing again 1170944478 M * daniel_hozac remove the chcontext. 1170944481 M * daniel_hozac does it still work? 1170944494 M * pflanze no, empty output 1170944507 M * daniel_hozac hum. 1170944517 M * daniel_hozac does removing the ncontext work? 1170944533 M * pflanze no, empty output, too. 1170944564 M * pflanze it requires ncontext AND chcontext AND 1005 AND nesting chcontext inside ncontext. 1170944584 M * daniel_hozac that is peculiar... 1170944600 M * pflanze ncontext --nid 1005 --migrate chcontext --ctx 1 lsof -i doesn't work either 1170944887 M * daniel_hozac ah, do you have CONFIG_VSERVER_PRIVACY enabled? 1170945002 M * pflanze yes 1170945073 M * pflanze Hm what has VSERVER_PRIVACY been created for? 1170945108 M * pflanze I've already started writing a script looping over vserver-stat output, so I could live with the above problem. 1170945112 M * daniel_hozac to make guests private ;) 1170945128 M * pflanze Private from ctx 1? 1170945134 M * daniel_hozac yep. 1170945166 M * pflanze ah, is that all? That should be better explained in the config help 1170945168 M * daniel_hozac with the privacy enabled, you have to enter the guest to get the output desired. 1170945921 M * pflanze ping 1170945943 M * pflanze Hm, I've got an oops 1170945956 M * pflanze again. 1170945970 M * pflanze 2.6.19.3-vs2.2.0-rc11 1170946160 M * pflanze http://paste.linux-vserver.org/1116 1170946377 M * pflanze Bertl_zZ: that looks to be the same as http://paste.linux-vserver.org/1104; looks like your guess that this is solved with rc11 doesn't hold true. 1170946452 M * pflanze oh, delta-cowbl-fix01.diff is not in rc11. 1170946501 M * pflanze So it has just been the wrong release you've mentioned. 1170946607 M * daniel_hozac we thought it was in rc11. 1170946613 M * daniel_hozac it was created against rc10, IIRC. 1170946618 M * daniel_hozac but we forgot it. 1170946622 M * daniel_hozac it's definitely in rc12 though. 1170946626 M * daniel_hozac (as the changelog says :)) 1170946676 Q * pflanze Read error: Connection reset by peer 1170946723 J * pflanze ~chris@84-73-56-44.dclient.hispeed.ch 1170946876 M * pflanze ok, Bertl_zZ: vs2.2.0-rc11 does not contain delta-cowbl-fix01.diff; and yes, delta-cowbl-fix01.diff on top of rc11 does seem to fix the http://paste.linux-vserver.org/1116 oops. 1170946921 M * daniel_hozac -rc12 contains delta-cowbl-fix01 and lots more. 1170947007 M * daniel_hozac (as per http://linux-vserver.org/ChangeLog-2.2) 1170947039 J * hiaslboy ~hiaslboy@85.127.227.37 1170947480 Q * chand Quit: chand 1170949581 Q * orzel Remote host closed the connection 1170950132 J * bXi_ bluepunk@irssi.co.uk 1170950133 Q * bXi Read error: Connection reset by peer 1170950558 Q * cdrx Quit: Leaving 1170950574 J * cdrx ~legoater@blueice1n1.uk.ibm.com 1170950701 M * daniel_hozac pflanze: the output of the vpkg script is processed by vhashify to create the exclude lists. 1170950709 M * daniel_hozac _that's_ why you don't see it. 1170950712 M * pflanze ah. 1170950720 M * daniel_hozac try stopping your guest before hashifying. 1170950724 M * pflanze yeah, I've just verified that non-running vservers cannot be hashified. 1170950745 M * pflanze so I'll change that on the wiki 1170950767 M * daniel_hozac that's only true for internal package management though. 1170950784 M * pflanze hm, but *when* is it being run? 1170950791 M * daniel_hozac when's what being run? 1170950796 M * pflanze vpkg 1170950802 M * daniel_hozac when you run vhashify. 1170950816 M * pflanze why would I run vhashify? 1170950821 M * pflanze or when 1170950832 M * daniel_hozac vserver ... hashify just calls vhashify. 1170950851 M * pflanze hm, so is the echo added to the eclude list? 1170950855 A * pflanze feeling dumb 1170950875 M * daniel_hozac yes. 1170950882 M * pflanze okay. 1170950941 A * pflanze retries with echo .. >&2 1170951002 M * pflanze I wonder: is "vserver .. enter" safe? IIRC there has been some general incertitude about this 1170951022 M * daniel_hozac safe in what way? 1170951046 M * pflanze That hijacked commands cannot get control of the host. 1170951057 M * pflanze through signals or whatever. 1170951086 M * pflanze iirc "vserver ... start" has been considered safe, "vserver .. enter" not. 1170951101 M * daniel_hozac AFAIK it's safe. 1170951128 M * pflanze That discussion has been more than a year ago. Maybe people trust in the meantime. 1170951179 M * pflanze Probably Bertl wanted to say back then, please verify the code, noone has done it, I won't take on responsibility. 1170951867 Q * transacid Remote host closed the connection 1170952011 J * chand ~chand@m167.net81-64-156.noos.fr 1170952029 J * FireEgl ~FireEgl@68.220.222.136 1170952744 Q * mountie Remote host closed the connection 1170952765 Q * kir Quit: Leaving 1170952769 J * mountie ~mountie@CPE0080c6fe323f-CM000a739acaa4.cpe.net.cable.rogers.com 1170953633 J * Aiken ~james@ppp102-23.lns1.bne1.internode.on.net 1170954645 J * stefani ~stefani@flute.radonc.washington.edu 1170954663 Q * chand Quit: chand 1170955559 M * pflanze Hm, how do you handle backups of unified vservers? 1170955564 M * pflanze hashified ones. 1170955593 M * pflanze When using rsync, I guess I have to backup the .hash directory as well, then upon restore, 1170955604 M * pflanze run over the files in .hash and restore the flags. 1170955662 M * pflanze Alternatively, restore each vserver from the backup separately. Then hashify again. 1170955709 M * pflanze Are there tools handling the immutable+unlinkable flags? 1170955723 M * pflanze dumpe2fs? 1170955730 M * daniel_hozac should work. 1170955742 M * pflanze of course not for reiserfs, I guess. 1170955760 M * daniel_hozac nope. 1170955776 M * daniel_hozac i just back up all of /vservers. 1170955807 M * pflanze Then restore flags in .hash/? 1170955872 M * pflanze ehr, as it seems dumpe2fs is not a backup utility, I must be mistaken in the name. 1170955970 M * pflanze "dump" is what I've meant. 1170955977 A * pflanze never used it 1170956357 J * bonbons ~bonbons@83.222.37.103 1170957966 M * pflanze Is there already a tool which can clone a vserver by taking advantage of hardlinks? 1170957982 M * pflanze instead of "cp -a" and running hashify afterwards. 1170958007 M * pflanze "cp --link" won't do, of course. 1170958045 M * pflanze Only files which already have the iunlink flags should be linked. 1170958201 M * hardwire rsync 1170958202 M * hardwire :) 1170958229 M * hardwire it will mimic cp -a 100% 1170958235 M * hardwire which means, use cp -a 1170958287 M * pflanze I'm talking about a tool which does not copy files which already have been hashified, but hardlink them. 1170958306 M * pflanze This would be faster. 1170958322 M * pflanze (And not use temporary space.) 1170958334 M * derjohn Radiance, Starting with 2.6.18 a user process must be root to manage tap interfaces. 1170958334 Q * cdrx Read error: Connection reset by peer 1170958380 M * derjohn Radiance, i just ran accoss it with qemu. My OpenVPN-i-a-guest box is older, that would maybe explain the cannot set mtu behaviour. 1170958383 M * daniel_hozac pflanze: vserver ... build -m clone. 1170958414 M * Radiance derjohn, thanks, i'm using 2.6.17.14 btw 1170958421 M * Radiance using a tun setup 1170958443 M * derjohn Radiance, that was just guess... but sounded logical ;) 1170958476 M * Radiance i've built a testbox and it's correct that one needs cap_net_admin (i saw some one else also battle this issue here) 1170958498 M * derjohn Radiance, OTOH the from the hosts point of view, all actions of the guest are 'done' by root. 1170958532 M * Radiance without cap_net_admin openvpn will never work in a guest since it will try to set the tx and mtu values regardless if root set them in the host 1170958563 M * derjohn Radiance, might true for tun mode. tap mode at least worked once ;) 1170958575 M * Radiance yeah tap mode is a different story hehe 1170958603 M * Radiance but with tap mode i would bridge a physical interface ?? 1170958633 M * Radiance or is it possible to bridge with tap mode eth0:1 for example ? (never tried tap mode so i'm just wondering here) 1170958648 M * derjohn Radiance, a bridge between remote and local. with brcontrol you can even bridge that onto the eth if you like. (of you need proxy_arp IIRC) 1170958729 M * Radiance nasty :) 1170958739 Q * dlezcano Read error: Connection reset by peer 1170958745 M * derjohn Radiance, cehteh should know more about it. 1170958752 M * Radiance but for a many clients - 1 server setup tun would be better ? 1170958792 M * Radiance ok thanks, i'll check with him for this too heh 1170958802 M * Radiance i hope to compile a nice doc which can be put in the wiki 1170958918 M * derjohn Radiance, good docs are always very welcome ! 1170958945 M * daniel_hozac derjohn: speaking of docs, what do you think of http://linux-vserver.org/Capabilities_and_Flags and http://linux-vserver.org/util-vserver:Capabilities_and_Flags? 1170958956 M * daniel_hozac (the separation idea, mostly) 1170959035 M * derjohn daniel_hozac, wow! 1170959047 M * derjohn daniel_hozac, thats even interesting for non-vserver ppl ! 1170959195 M * daniel_hozac well, i didn't write the first one, that's been there for a rather long time ;) 1170959225 M * daniel_hozac i'd like to use a similar scheme for all the pages though. 1170959260 M * derjohn daniel_hozac, thats an intensive thing to do so. 1170959284 M * daniel_hozac e.g. move the Configure Resource Limits in util-vserver section from http://linux-vserver.org/Resource_Limits to a http://linux-vserver.org/util-vserver:Resource_Limits page. 1170959316 J * transacid ~transacid@transacid.de 1170959385 A * derjohn must go. bye ppl, cu l8ter ! 1170959547 M * nox http://paste.linux-vserver.org/1118 <- anyone an idea to that? 1170959593 M * daniel_hozac interesting... do you have the lines above that? 1170959612 M * daniel_hozac (the ones that say what type of problem it is 1170959617 M * nox that was pastet on all terms, sec search the logs 1170959904 M * nox no nothing :( 1170959991 M * daniel_hozac how did you get it? 1170960291 M * nox found it http://paste.linux-vserver.org/1119 1170960311 M * nox was getting big sensors alarm this night 1170960506 M * daniel_hozac do you still have the source tree? 1170960519 M * daniel_hozac could you run addr2line -e vmlinux c0140877? 1170960580 M * nox addr2line: 'vmlinux': No such file 1170960623 M * daniel_hozac inside your kernel tree? 1170960639 M * nox ah doh 1170960673 M * nox ??:0 1170960691 M * daniel_hozac so i guess grep CONFIG_DEBUG_INFO .config returns not set? 1170960790 M * nox doesn?t even have this option in my conf 1170960796 M * daniel_hozac okay :) 1170960837 M * nox ok gonna turn it on again next time 1170960839 Q * michal`_ Ping timeout: 480 seconds 1170960872 M * dm8tbr ---------------------- 1170960873 M * dm8tbr ---------------------- 1170960876 M * dm8tbr ---------------------- 1170960909 M * nox dm8tbr: you are here? 1170960924 M * nox ltns 1170961383 J * michal` ~michal@www.rsbac.org 1170961413 M * nox another question,any bind9 issues with >2.1.1-rc38? 1170961487 M * nox i have 2 guest with bind9 running without any prob , but couldn?t get an answer from one guest i got provided by transacid 1170961558 M * nox only difference i can see is that my bound to puplic ips and his one prerouted to a nonpublic 1170961585 M * nox zbut that shouldn?t make a difference or? 1170961787 M * nox when he started bind9 on the host it worked 1170962375 M * daniel_hozac depends on your rule, i guess. 1170962393 M * daniel_hozac BIND is pretty picky about the source IP addresses. 1170962497 M * nox so its more a nat-bind than a vserver prob? 1170962509 M * daniel_hozac no idea... 1170962528 N * DoberMann[PullA] DoberMann 1170962567 M * nox it opens a socket but even dig on the local ip doesn?t work 1170963097 Q * michal` Ping timeout: 480 seconds 1170963646 J * michal` ~michal@www.rsbac.org 1170964934 M * dm8tbr nox: whoops. :) 1170965513 N * bXi_ bXi 1170966744 Q * lilalinux Remote host closed the connection 1170968004 M * pflanze daniel_hozac: is "vserver ... build -m clone" using some simpler tool underneath? 1170968012 M * pflanze I like simple tools. 1170968014 M * daniel_hozac vclone. 1170968062 M * pflanze hm no vclone in my installation 1170968094 M * daniel_hozac it was added in 0.30.213-pre4. 1170968094 M * pflanze daniel_hozac: ah, vcopy? 1170968103 M * daniel_hozac no. 1170968114 M * pflanze ah, this is 0.30.212 1170968522 M * pflanze daniel_hozac: where can I get 0.30.213* ? 1170968530 M * daniel_hozac http://people.linux-vserver.org/~dhozac/p/uv/experimental/ 1170968669 M * pflanze sha1sum 275a1406d8acf8cbfcd00f3170c77b215b2e6f02 util-vserver-0.30.213-rc1.tar.bz2 1170968757 M * daniel_hozac yep. 1170969198 Q * dm8tbr Ping timeout: 480 seconds 1170969445 M * pflanze two small notes: vclone w/o any argument gives "vclone: chdir(): Bad address" and giving a relative --dest doesn't work because of chdir. 1170969467 M * pflanze would you like patches? 1170969471 M * pflanze or is this in flux anyway? 1170969556 M * daniel_hozac --dest and --source will be going away. 1170969866 M * xp_prg2 is it possible to duplicate a vserver with a differnet hostname and ip adress etc... ? 1170969983 M * daniel_hozac that's what -m rsync and -m clone do. 1170970296 M * xp_prg2 ok thanks I will try that 1170970439 J * dm8tbr ~dm8tbr@dk0td.afthd.tu-darmstadt.de 1170970864 M * cehteh daniel_hozac: did you read about my strange vserver problem in the moring? 1170971232 M * daniel_hozac and replied ;) 1170971247 M * daniel_hozac 13:27 < daniel_hozac> cehteh: you have the legacy version enabled. that will disable the new chbind, and if you don't build the tools with legacy support, there's nothing to fallback to 1170971278 M * cehteh ahok 1170971310 M * cehteh heh ... my nick was highlighted so i scrolled back to see why ... 1170971320 M * cehteh [19:19] Radiance, cehteh should know more about it. 1170971327 J * Piet_ hiddenserv@tor.noreply.org 1170971328 M * cehteh .. G* so i missed your answer 1170971346 M * cehteh derjohn: spammer :P 1170971485 M * Radiance cehteh, well we're discussing openvpn in tun mode in a vserver, it'll work but only with cap_net_admin (as far as we can see:) 1170971523 M * Radiance since it has to set the mtu, tx values and up routes 1170971528 M * Radiance without this it won't work 1170971583 M * cehteh i played once with it ... the trick is to setup the tun/tap as persistent with tunctl (or how it was named) in the host ... tunctl is part of uml-utils 1170971601 M * Radiance done already 1170971629 M * Radiance won't work since it'll need evil powers to manipulate the interface settings 1170971644 M * Radiance at least not for tun mode 1170971647 M * Radiance i didn't try tap mode 1170971649 M * cehteh well .. next i have a shitload of briging problems to solve .. at some point i had it working but i lost interest since it was not fully stable 1170971664 M * Radiance yeah i can imagine hehe 1170971678 M * cehteh and i tried a layer2 vpn with the new ssh tunneling features not openvpm 1170971689 M * Radiance well at least the vps is owned by myself so it's not that risky, it just means thirdparties cannot get this kind of power 1170971742 M * Radiance yeah ssh tunneling is nice, but it won't automagically redirect all traffic heh 1170971757 M * daniel_hozac OpenSSH can do VPNs now. 1170971760 Q * Piet Ping timeout: 480 seconds 1170971762 M * cehteh yes 1170971774 M * cehteh but there is quite a lot to do manually 1170971785 M * daniel_hozac yeah. 1170971798 M * cehteh Radiance: not port redirection .... real tun/tap tunneling with ssh 1170971856 M * cehteh daniel_hozac: what would you suggest to fix my problem .. rebuild kernel without legacy support or how do i build the tools with legacy support? 1170971900 M * daniel_hozac kernel without the legacy version would probably be the best idea, as util-vserver would use the new APIs. 1170971907 M * cehteh ok 1170971915 M * daniel_hozac but if you want to rebuild the utils, --enable-apis=NOLEGACY is the key. 1170971925 M * cehteh disable both legacy things (kernel and network) 1170971958 M * cehteh i just rebuild the kernel 1170971974 M * daniel_hozac the legacy version is the weird option. 1170971982 M * daniel_hozac why did you enable that in the first place? 1170972050 M * cehteh ages ago i had the opposite problem ;) 1170972076 M * cehteh and usually enabling legacy deprecated thing is the safe side in linux 1170972077 M * daniel_hozac did you use the ancient utils? 1170972087 M * daniel_hozac not the legacy version. 1170972092 M * cehteh yes ... very long ago ... 1170972151 M * cehteh ok building 1170972427 M * Radiance cehteh, interesting indeed 1170972442 M * Radiance i will check that out 1170972496 M * cehteh Radiance: uhm .. i think you want openvpn for serious tunneling 1170972547 M * Radiance well, i'm still curious to the ssh part, can't hurt to play with that too 1170972682 M * Radiance i remember derjohn saying that util-vserver 310 and below had a bug causing problems with openvpn routing ? 1170972707 M * Radiance does this affect tun mode in a server - clients situation ? 1170972725 M * Radiance i believe daniel_hozac discovered it 1170972751 M * daniel_hozac the peer problem was fixed in 2.0.2-rc22 or so 1170972784 M * daniel_hozac and personally, i don't consider it a bug that util-vserver doesn't know how to configure a peer. 1170972988 Q * DavidS Quit: Leaving. 1170973255 Q * ema Quit: 'night 1170973279 Q * bonbons Quit: Leaving 1170973426 N * DoberMann DoberMann[ZZZzzz] 1170974060 N * Bertl_zZ Bertl 1170974065 M * Bertl morning folks! 1170974158 M * daniel_hozac morning Bertl! 1170974373 M * Radiance daniel_hozac, cool then it should be fine in 2.0.2.1 :) 1170974456 M * harry linux-image-generic 1170974460 M * harry and linux-headers-generic 1170974469 M * harry both depend on linux-headers-2.6.17-11-generic 1170974476 M * harry but... that package does not exist??? 1170974481 M * harry how is that possible??? 1170974488 M * Bertl Provides: ? 1170974491 M * harry another ubuntu hippy-feature? 1170974505 M * harry Bertl: ? 1170974527 M * Bertl at least in rpm there is the 'Provides' statement 1170974530 M * harry there is no 2.6.17-11!!! 1170974533 M * Bertl (rpm/spec) 1170974536 M * harry it doesn't exist 1170974541 M * harry nowhere in the world :S 1170974556 M * Bertl well, it can be a bug as well, I do not know ubuntu 1170974586 M * harry damn... 1170974589 M * daniel_hozac completely OT question: anyone know if it's possible to override a PHP function, and maybe call the real one? 1170974609 M * daniel_hozac sort of like #define eval my_eval would do in C? 1170974633 M * harry i'll look somewhat further 1170974691 M * ntrs Bertl, is there a 2.2.0-rc12 for 2.6.20? 1170974704 M * daniel_hozac no. 2.6.20 has -pre3. 1170974711 M * ntrs ah, I see 1170974726 M * ntrs actually no, 2.6.20 is out. 1170974733 M * ntrs Just checked. 1170974763 M * daniel_hozac yes, but what i meant was that 2.2.0-pre3 is available for 2.6.20. 1170974793 M * ntrs daniel_hozac, isn't pre3 newer than rc12? 1170974807 M * daniel_hozac yes. 1170974816 M * daniel_hozac but the branches are disjoint. 1170974824 M * ntrs huh? 1170974843 M * ntrs is 2.2.0-rc3 equivalent to 2.2.0-rc12 with some fixes? 1170974879 M * ntrs I meant 2.2.0-pre3 1170974908 M * daniel_hozac since it's against another base kernel, they're not equivalent at all. 1170974915 Q * Wonka Read error: Connection reset by peer 1170974928 M * daniel_hozac 2.6.19.3-vs2.2.0-rc12 is the latest in the 2.6.19-vs2.2 branch. 1170974941 M * daniel_hozac 2.6.20-vs2.2.0-pre3 is the latest in the 2.6.20-vs2.2 branch. 1170974951 M * daniel_hozac separate branches, same/similar vserver code. 1170974956 M * ntrs so, are you saying that there could be some features that are in rc12 but are not in pre3? 1170974962 J * Wonka produziert@chaos.in-kiel.de 1170974980 M * daniel_hozac i haven't checked the deltas for the 2.6.20 branch, so i don't know. 1170974991 M * ntrs Bertl, ? 1170975121 M * daniel_hozac -pre2 seems to have all the -rc12 changes. 1170975132 M * daniel_hozac -pre3 have some further reaper fixes. 1170975160 M * ntrs ok, is 2.2.0-pre3 at least as stable as rc12? 1170975168 M * Bertl ntrs: nope, 2.6.20 has everything 1170975182 M * daniel_hozac i doubt that. 1170975187 M * Bertl ntrs: but 2.6.20 is basically untested, so don't use it in production 1170975190 M * daniel_hozac 2.6.19 is just arting to stabilize... 1170975199 M * ntrs Ok 1170975200 M * daniel_hozac +st 1170975203 Q * dna Quit: Verlassend 1170975205 M * pflanze daniel_hozac: what's this?: vserver imap hashify -> sh: /lib/tls/libc.so.6: version `GLIBC_2.3.4' not found (required by sh)\n failed to determine configfiles 1170975221 M * pflanze And it stops. When I try again, the error doesn't happen anymore. 1170975237 M * Bertl pflanze: missing library for your shell? 1170975261 M * ntrs Can someone tell me if an existing running host with multiple running guests can be unified (hashified?) and start using COW? what is needed? 1170975290 M * Bertl CoW Link Breaking works mostly behind the scenees 1170975291 M * daniel_hozac you'd have to run vserver ... hashify and restart the guests. 1170975293 M * pflanze This is the second new vserver I'm seeing this now. It's always upon the first hashification of a new vserver it seens. And it only happens since I've upgraded to 30.213. 1170975314 M * ntrs Bertl, daniel_hozac, is there some easy howto or some other document? 1170975350 M * daniel_hozac http://www.paul.sladen.org/vserver/archives/200512/0323.html 1170975352 M * pflanze Bertl: how would that happen? sh on the host doesn't have a problem, sh on the client neither. 1170975388 M * ntrs I assume there needs to be a reference template/image that all the other guests of the same distro will point to, right? 1170975389 M * pflanze Isn't this an issue with loading of shared libs of a process started from the host in client context? 1170975400 M * Bertl ntrs: nope 1170975418 M * Bertl ntrs: this will be created in the cache automatically 1170975421 M * ntrs Bertl, how? What will the hardlinks pont to? 1170975428 M * ntrs what cache is that? 1170975429 M * daniel_hozac the .hash. 1170975452 M * Bertl pflanze: hmm, could be some obscure effect with preloaded glibc libraries 1170975465 M * Bertl pflanze: IIRC, vhashify still uses glibc 1170975492 M * Bertl daniel_hozac: btw, from yang: 1170975497 M * ntrs Is all this hashify/unify/cow well tested and stable? 1170975500 M * Bertl 21:14 2 of 16 tests failed 1170975500 M * Bertl 21:14 Please report to vserver@list.linux-vserver.org 1170975506 M * Bertl on sparc32 1170975509 M * daniel_hozac hashify? 1170975514 M * daniel_hozac on Debian? 1170975522 M * Bertl debian, yes 1170975537 M * Bertl is there a known fix for that? 1170975538 A * cehteh uses hashify on debian too 1170975584 M * pflanze ntrs: I'd say depends. The above error makes me scare. 1170975603 M * daniel_hozac the Debian people would have to answer that... i think the problem is with beecrypt2. 1170975608 M * Bertl ntrs: Aiken has done a lot of testing, and IIRC, his guest hoppy is 100% hashified 1170975632 M * daniel_hozac ntrs: hashify/unify is just the userspace. it just links/copies the files. 1170975650 M * Bertl daniel_hozac: can we drop beecrypt sooner or later and switch to openssl maybe? 1170975668 M * Bertl daniel_hozac: or even better, include a custom hash algo? 1170975685 M * Bertl (and have that built with dietlibc) 1170975697 M * daniel_hozac GNU TLS maybe. 1170975699 M * ntrs daniel_hozac, Bertl, does a guest need to be hashivied/unified only once, and never again? 1170975717 M * Bertl daniel_hozac: I can immediately providde various hash algos if that is the problem ... 1170975722 M * daniel_hozac ntrs: the hashification needs to happen at regular intervals, as updates would unlink the files. 1170975738 M * ntrs daniel_hozac, what updates? yum? apt-get? 1170975750 M * pflanze ntrs: yes. 1170975764 M * Bertl ntrs: typically you do the hashification periodically or when you think that you need some space :) 1170975767 M * ntrs daniel_hozac, what is a normal/default interval? 1170975774 M * daniel_hozac ntrs: there is none. 1170975780 M * ntrs Ok 1170975785 M * Bertl ntrs: there is no harm in _not_ doing so :) 1170975787 M * daniel_hozac i do it every week from my yum cron job. 1170975810 M * pflanze ntrs: see my new wiki page on http://linux-vserver.org/util-vserver:Vhashify 1170975819 M * ntrs I think this whole thing looks quite complex and complicated. 1170975820 M * daniel_hozac Bertl: i guess the problem isn't much with the algorithm, but rather that it needs to be versatile. 1170975832 M * daniel_hozac i.e. support multiple algorithms... 1170975836 M * xp_prg2 No X11 DISPLAY variable was set, but this program performed an operation which requires it. 1170975866 M * pflanze xp_prg2: configured sshd to allow X? 1170975872 M * daniel_hozac ntrs: if you just want COW from a template, that's easier. 1170975891 M * ntrs yes, cow from a template should be enough for me. 1170975923 M * daniel_hozac ntrs: just find /vservers/