1170806499 Q * hiaslboy Ping timeout: 480 seconds 1170807189 Q * orzel Read error: Connection reset by peer 1170807193 J * orzel ~orzel@freehackers.org 1170807309 M * orzel PowerKe: ok, i'm running -rc11 now 1170807313 M * orzel but i have the same problem :-(((((((( 1170807341 M * orzel i'm re-installing util-vserver...just in case 1170807350 M * PowerKe Probably something with the guest init-scripts then 1170807384 M * orzel how can i do then ? 1170807393 M * orzel i've already removed the 'style' file for apps/init 1170807434 M * orzel No command given; use '--help' for more information. <---- vserver fhweb start says that 1170807439 M * orzel i wonder where this comes from 1170807449 M * PowerKe I'm thinking about the /etc/inittab /etc/init.d/* stuff in your guest 1170807462 M * PowerKe I think you better put the plain style back 1170807478 M * orzel ok, but it doesn't print anything neither 1170807495 M * PowerKe It never does 1170807519 M * orzel ah, ki 1170807521 M * orzel ok 1170807524 M * orzel now it says 1170807525 M * orzel vcontext: jailIntoTempDir(): No such file or directory 1170807525 M * orzel vcontext: write(): Broken pipe 1170807555 M * orzel gg doesn't know about this 1170807647 M * orzel util-vserver refuses to get emerged 1170807654 M * PowerKe I guess you could try to emerge vserver-baselayout into the guest 1170807657 M * orzel it looks like the end of the world is not far enough 1170807669 M * PowerKe what does util-vserver complain about? 1170807685 M * orzel PowerKe: i used to do that from within the vserver. i know there's gentoo script to do that from host 1170807717 M * PowerKe ROOT="/vservers/yourgest" emerge -pv vserver-baselayout 1170807725 M * orzel says that : http://rafb.net/p/VvPwOx61.html 1170807743 M * PowerKe and that's baselayout-vserver 1170807768 M * orzel lol, indeed :) 1170807828 M * PowerKe sys-cluster/util-vserver-0.30.212-r2 fails here too 1170807833 M * PowerKe Try -r1 1170807845 M * orzel i was just doing that :) 1170807863 Q * Greek0_ Ping timeout: 480 seconds 1170807961 M * orzel PowerKe: how do i etc-update in the host? 1170807970 M * orzel same way ROOT=? 1170807977 M * PowerKe chroot 1170807980 M * orzel k 1170808021 M * orzel ok, it's done, but i've got the same error trying to start the vserver :( 1170808071 M * orzel actually, i want to find again util-vserver-0.30.211 1170808073 M * orzel this one works 1170808086 M * orzel but all my computer are synce'd, and they've removed it from portage 1170808111 M * orzel strange, btw, there's only two ebuilds for util-vserver. while for other packages there are ~8/10,and quite old 1170808126 M * PowerKe services I have in my guests (apart from apache and stuff): boot level: bootmisc, domainname, hostname, rmnologin 1170808133 M * PowerKe default level: local 1170808186 M * orzel i have that : apache2 local syslog-ng 1170808193 M * orzel (default) 1170808195 M * orzel and that : 1170808195 M * orzel bootmisc domainname hostname rmnologin 1170808199 M * orzel for 'boot' 1170808202 M * orzel which seems ok to me 1170808211 M * PowerKe looks good 1170808319 M * PowerKe Did you remove fstab and modprobe.conf? 1170808341 M * PowerKe and add fstab to /etc/vservers/yourguest ? 1170808454 M * orzel sure, it worked until i used 112 1170808469 M * orzel i've got a bit fstab in /etc/vservers/guest 1170808475 M * orzel i've even strimmed it down 1170808480 M * orzel (to check if it prevents booting) 1170808482 M * orzel but no 1170808506 M * orzel actually i used gentoo stuff to create the guest, so there was no fstab in guest/etc/ 1170808564 M * orzel i've found a 111-rc1 1170808568 M * orzel going to try that 1170808603 M * PowerKe vserver guest start --rescue and then from another console vserver guest enter ? 1170808635 M * orzel ohhh 1170808639 M * orzel that seems to work 1170808648 M * PowerKe From what I can tell it seems like --rescue starts a sleep 900 inside the guest, which keeps the guest running for a while 1170808680 M * PowerKe You should start at least 1 service inside the guest, but you have both syslog-ng and apache2 so that should work 1170808694 M * orzel i can even start apache 1170808709 M * orzel which works, i can browse 1170808721 M * orzel syslog as well 1170808737 M * PowerKe But there still has to be something wrong in the startup scripts 1170808753 M * PowerKe Right now your guest is going to shut down again in 15 minutes (try ps -ef) 1170808792 M * PowerKe Unless it would keep on running because you started apache now... 1170808827 M * orzel anyway i have to find out why it wont start the normal way 1170808830 M * orzel how can i do ..? 1170808833 M * orzel everything seems fine 1170808900 M * orzel there's indeed a 'sleep 900' 1170808995 M * orzel all services in boot/ and default/ can restart without error 1170809039 M * orzel all is mounted right. this is not the pb 1170809104 M * PowerKe Maybe check one of the services' startup files to see if the guest fails before or after starting the service 1170809119 M * PowerKe (apache log or syslog) 1170809131 M * orzel apache/syslog weren't started 1170809138 M * orzel even if rc-status says so 1170809176 M * PowerKe You'd have to check with a normal vserver start, I think that --rescue bypasses the entire init script 1170809238 M * orzel what can i put instead of 'plain' to see what happens ? 1170809247 M * PowerKe gentoo ? 1170809263 M * orzel no, i haven't the proper baselayout 1170809268 M * orzel Using init-style 'gentoo' requires >=baselayout-1.13 inside the vserver! 1170809274 M * orzel (it says) 1170809309 M * PowerKe hmm, hollow pasted an url earlier that would allow the gentoo style with older baselayout 1170809312 M * orzel mmmmm, i can't stop the vserver now :( 1170809348 M * PowerKe just kill the vserver guest start --rescue ? 1170809366 M * PowerKe hmm, maybe apache keeps it running now 1170809391 M * orzel i've ^C it 1170809396 M * orzel but it stays here 1170809401 M * orzel hope the 15min stuff will quit 1170809414 M * PowerKe I think it keeps on running while something is attached to the init of the guest 1170809424 M * PowerKe Since you started apache & syslog, they keep it running 1170809434 M * PowerKe can you enter the guest and stop the services? 1170809551 M * orzel okkkkkkkkkkkkkkk 1170809553 M * orzel i've fixed it 1170809566 M * orzel 1) i removed apache/syslog, and the vserver stopped 1170809576 M * orzel 2) the problem is that /tmp on my HOST is a link 1170809587 M * orzel creating a local, non-link /tmp solved it all 1170809594 M * orzel i'm just VERY surprised about that 1170809605 M * PowerKe How did you figure that out? 1170809670 M * orzel the msg speaks about tmp 1170809690 M * orzel the one i copy/pasted here 1170809701 M * orzel pb is / has not enough space to handle /tmp 1170809708 M * PowerKe right, still, nice catch 1170809741 M * orzel vcontext: jailIntoTempDir(): No such file or directory 1170809742 M * orzel was the msg 1170809752 M * orzel i confirm, i switched back to /tmp being a link and it fails 1170809759 M * dm8tbr how about mount --bind 1170809784 M * orzel i think util-vserver needs to be fixed, not my system ! :-) 1170809788 M * orzel but meanwhile, yes, why not 1170809808 M * orzel although the link was against some dir on some other partition, not a partition itself 1170809811 M * orzel it will work ? 1170809812 M * PowerKe or use tmpfs if you have enough ram/swap 1170809812 M * orzel let's try 1170809821 M * orzel i'm short with ram 1170809822 M * orzel only 1gig 1170809825 M * dm8tbr orzel: thats what the --bind is for 1170809831 M * orzel with 1 host and 2 vservers 1170809833 M * orzel dm8tbr: ok:) 1170809847 M * dm8tbr it binds a directory to another one 1170809871 M * dm8tbr dunno how good / bad that might work 1170809899 M * orzel it works 1170809900 M * PowerKe Mounting tmpfs on /tmp works, so I suppose --bind should work as well 1170809916 M * orzel it does:) 1170809924 M * orzel how do I say the same thing in fstab ? 1170809992 M * PowerKe /tmp /target none bind,rw 0 0 1170810005 M * orzel ok, found 1170810008 M * PowerKe (that's how it is in /etc/vservers/guest/fstab :) ) 1170810018 M * orzel without the rw, but it's default anyway 1170810037 M * orzel pfffff 1170810041 M * orzel 2:0 am 1170810048 M * orzel this was quite a long one :) 1170810053 M * PowerKe oh, you're in my timezone :) 1170810056 N * DreamerC_ DreamerC 1170810060 M * orzel your .be right ? 1170810063 M * orzel i'm .fr 1170810064 M * PowerKe idd 1170810079 A * orzel in Paris 1170810099 A * PowerKe near Antwerp 1170810107 M * orzel let's have a biiiiig cold water glass now :) 1170810108 M * PowerKe Anyways, going to bed now 1170810112 M * orzel so am i 1170810134 M * orzel PowerKe: thanx a lot for the time we spent together. was a lot less boring that fighting alone with the bug :) 1170810145 M * orzel dm8tbr: thx for the --bind hint :) 1170810155 M * dm8tbr np :) 1170810164 M * PowerKe np, nice to learn something more 1170810197 M * orzel i've always had this /tmp link 1170810205 M * orzel util-vserver 111 was not using /tmp the same way 1170810244 M * PowerKe ls -al /tmp doesn't show anything vserver related when the vserver is running 1170810253 M * PowerKe Maybe it just needs it when starting though... 1170810266 M * orzel yes 1170810277 M * orzel start/stop script creates a tmp dir with a makefile that is executed 1170810294 M * orzel and remove the directory after that 1170810331 M * PowerKe See, I'm learning a lot more tonight than I'm actually helping :) 1170810391 M * orzel i've read all those scripts to try to find the pb :).. 1170810438 M * PowerKe hmm, on my old vserver I do have vserver-init. files in /tmp 1170810477 M * PowerKe utils = .210, kernel = 2.0.1 1170812742 J * Greek0 ~greek0@85.255.145.201 1170813300 Q * Aiken Quit: Leaving 1170814642 Q * SNy Ping timeout: 480 seconds 1170814877 M * litage one of my vserver guests can't bind to its ip address. ``ifconfig'' and ``ifconfig -a'' output nothing. when stopping the guest, this error occurs: "Rebooting... RTNETLINK answers: Cannot assign requested address" 1170814884 J * SNy 3f9069a9e5@bmx-chemnitz.de 1170814919 M * litage running ``ip addr'' on the vserver host does not list the ip address the problem-guest is supposed to use. any suggestions on how to debug or fix this? 1170816016 Q * thunder1 Remote host closed the connection 1170816031 J * thunder1 ~thu@tor-irc.dnsbl.oftc.net 1170816180 M * cehteh litage: too long name for the interface alias prolly 1170816536 M * litage cehteh: not sure what you mean.. 1170816575 M * litage cehteh: the vserver guest was working fine for 6+ months. it just suddenly fell over now 1170816580 M * cehteh yes 1170816587 M * cehteh whats the interface name? 1170816604 M * cehteh cat /etc/vservers/*/interfaces/*/name 1170816611 Q * thunder1 Remote host closed the connection 1170816625 J * thunder1 ~thu@tor-irc.dnsbl.oftc.net 1170816643 M * cehteh i had that too, seems the possible length got reduced with newer kernels 1170816807 M * litage cehteh: /etc/vservers/*/interfaces/*/name doesn't exist (i'm running a very old version): http://rafb.net/p/k2pEUe96.html 1170817007 M * cehteh then read about the new config and double check what could be wrong ... i cant tell from here 1170817030 M * cehteh did you updated vserver-util too? 1170817170 M * litage cehteh: nothing's been updated 1170817203 M * cehteh you updated the kernel? 1170817214 Q * thunder1 Remote host closed the connection 1170817225 J * thunder1 ~thu@tor-irc.dnsbl.oftc.net 1170817230 M * litage cehteh: nope 1170817257 M * cehteh huh .. well somehow it must become broken or? .. updated initscripts whatever 1170817278 M * litage cehteh: nothing's been updated on the box in months 1170817295 M * cehteh did you try a reboot? 1170817315 M * litage cehteh: i haven't rebooted the physical machine yet, cuz i'd prefer to solve this "properly" 1170817318 M * cehteh the guest worked before? 1170817355 M * cehteh check with vps aux | less is there is still something running in the context 1170817401 M * cehteh if yes kill that ... or maybe just try a reboot and check if the error shows up again 1170817416 M * litage cehteh: yeah, the guest's been working for 6+ months. it just fell over the other day 1170817486 M * cehteh maybe some zombie or whatever ... maybe even some bug dunno ... some servers (like squid) take quite long to shutdown you have to fix the vserver timeout 1170817499 M * cehteh could be anything .. and i cant tell you 1170817740 M * litage cehteh: i can't run vps: vc_new_s_context(): Operation not permitted 1170817752 M * litage cehteh: ``ps auxf'' doesn't list any zombies 1170817762 M * cehteh from the root server 1170817766 M * litage yeah 1170817776 M * cehteh huh strange 1170817792 M * cehteh double check that you are in the root server 1170817801 M * cehteh as root 1170817804 M * litage cehteh: trust me, i am :) 1170817831 M * cehteh mhm swallowed a blue pill? :) 1170817854 M * litage heh 1170817867 M * litage i guess i'll reboot the box.... =/ 1170817872 M * cehteh yeah 1170818411 M * litage cehteh: rebooting the physical machine doesn't seem to have helped 1170818433 M * cehteh mhm 1170818442 M * litage cehteh: the problem-guest still hasn't bound its ip address, and running ``ip addr'' on the vserver host doesn't list the problem-guest's ip address 1170818480 M * cehteh dunno .. maybe you wait for someone who knows vserver better than me 1170818490 M * litage thanks for your help cehteh 1170818522 M * litage has Bertl been around lately? 1170818523 M * cehteh i would start it under strace supervision ... 1170818556 M * cehteh dunno i just looked ... but its 4:22 am in europe .. he prolly wakes up soon ;) 1170818634 M * litage cehteh: like this?: strace -f -t -o vserver.strace vserver golive start 1170818811 M * litage nope, that didn't seem to work.... 1170818861 M * cehteh at least worth a try ... prepare for a big log 1170818955 M * litage cehteh: is /etc/vservers/NAME/namespace supposed to exist? 1170818994 M * cehteh nope 1170819010 M * cehteh at least i dont have it either ... defaults to NAME then 1170819032 M * cehteh you have /context ? 1170819062 M * cehteh with recent kernels thats mandatory 1170819102 M * litage /context/ doesn't exist on the vserver host or the problem guest 1170819159 M * cehteh i meant /etc/vserver/NAME/context 1170819177 M * cehteh vservers 1170819194 M * cehteh well if it worked before and you didnt changed the kernel, that isnt the problem 1170819588 J * _dmax ~semaj@81.193.56.237 1170819720 M * litage cehteh: /etc/vservers/*/context doesn't exist 1170819723 M * litage for me 1170819758 N * Bertl_oO Bertl 1170819775 M * Bertl litage: well, you should add it then :) 1170819805 M * litage hey hey, Bertl! been a long time :) 1170819825 M * litage Bertl: i'm running a very old version, and can't upgrade atm 1170819868 M * litage what should be in /etc/vservers/*/context ? 1170819879 M * Bertl the static context id 1170819888 M * Bertl (number between 2 and 49151) 1170819926 Q * dmax Ping timeout: 480 seconds 1170819932 M * Bertl okay, was a long day ... have to hit the bed now ... 1170819934 N * _dmax dmax 1170819949 M * Bertl have fun! cya tomorrow ... 1170819949 M * litage Bertl: i don't believe i've ever had this 'context' file, and up until just a day or two ago, all of my vserver guests were running fine... 1170819955 M * litage cya Bertl 1170819965 M * Bertl and now they are not? 1170819993 M * Bertl what is the misbehaviour? 1170820150 M * Bertl okay, sorry, off now .. good luck ... 1170820155 N * Bertl Bertl_zZ 1170822297 J * Aiken ~james@ppp220-70.lns2.bne1.internode.on.net 1170824484 Q * litage Ping timeout: 480 seconds 1170824526 Q * mcp Read error: Connection reset by peer 1170824571 J * mcp ~hightower@83.136.81.245 1170825120 Q * Aiken Quit: Leaving 1170825128 J * litage ~nick@203.220.55.70 1170826013 J * [3rdr00t thirdRoot@adsl-69-235-193-62.dsl.irvnca.pacbell.net 1170826259 M * [3rdr00t hey can someone help me in installing debian? 1170826292 N * [3rdr00t bob 1170826337 Q * bob 1170826597 J * bob thirdRoot@adsl-69-235-193-62.dsl.irvnca.pacbell.net 1170826603 M * bob can someone help me in installing debian? 1170826696 M * litage bob: wrong channel. try #debian 1170826873 M * bob they wont let me in 1170826937 M * litage bob: join #debian on FreeNode or EFNet 1170827208 Q * bob 1170827648 J * meandtheshel1 ~markus@85-124-37-28.dynamic.xdsl-line.inode.at 1170828026 J * rgl ~Rui@84.90.10.107 1170828093 N * DoberMann[ZZZzzz] DoberMann 1170828462 J * Aiken ~james@ppp220-70.lns2.bne1.internode.on.net 1170829856 Q * Greek0 Remote host closed the connection 1170829871 J * Greek0 ~greek0@85.255.145.201 1170830549 Q * orzel Read error: Connection reset by peer 1170830580 J * orzel ~orzel@82.225.154.2 1170832225 N * DoberMann DoberMann[PullA] 1170832941 J * Johnsie ~jdlewis@jdlewis.org 1170833009 Q * Johnnie Ping timeout: 480 seconds 1170833182 Q * orzel Read error: Connection reset by peer 1170833215 J * orzel ~orzel@82.225.154.2 1170833273 Q * cdrx Ping timeout: 480 seconds 1170833696 J * DoberMann_ ~james@AToulouse-156-1-38-142.w81-49.abo.wanadoo.fr 1170833802 Q * DoberMann[PullA] Ping timeout: 480 seconds 1170834139 Q * Aiken Quit: Leaving 1170834756 J * fosco_ fosco@konoha.devnullteam.org 1170834874 Q * fosco Ping timeout: 480 seconds 1170836184 N * DoberMann_ DoberMann 1170836245 J * benj0r ~benjamin@foxhound.sherpadown.net 1170836688 J * hiaslboy ~hiaslboy@85.127.227.37 1170836984 Q * ZLinux Remote host closed the connection 1170837025 J * cdrx ~legoater@blueice4n2.uk.ibm.com 1170837740 Q * benj0r Quit: Pwet 1170838858 N * fosco_ fosco 1170840019 J * Piet hiddenserv@tor.noreply.org 1170841224 J * dna ~naucki@153-227-dsl.kielnet.net 1170841345 Q * shedi Quit: Leaving 1170843659 Q * DoberMann Ping timeout: 480 seconds 1170844134 Q * infowolfe Read error: Connection reset by peer 1170844197 J * infowolfe ~infowolfe@c-67-164-195-129.hsd1.ut.comcast.net 1170844323 J * DoberMann ~james@AToulouse-156-1-50-132.w90-16.abo.wanadoo.fr 1170844409 J * shedi ~siggi@ftth-237-144.hive.is 1170844671 J * Zap-W Zap-W@87.69.58.137.cable.012.net.il 1170844676 P * Zap-W 1170844960 J * debugger ~Rui@84.90.10.107 1170845011 M * daniel_hozac hiaslboy: http://vserver.13thfloor.at/Experimental/patch-2.6.20-vs2.2.0-pre1.diff 1170845109 M * daniel_hozac hiaslboy: note that it's highly experimental at this point, and (AFAIK) largely untested. 1170845239 Q * rgl Ping timeout: 480 seconds 1170845331 J * duckx ~Duck@tox.dyndns.org 1170845481 J * ema ~ema@lart.galliera.it 1170846081 J * lilalinux ~plasma@80.69.41.2 1170846224 Q * cdrx Quit: Leaving 1170847214 Q * hiaslboy Ping timeout: 480 seconds 1170847348 J * DavidS ~david@vpn.uni-ak.ac.at 1170847519 Q * lilalinux Remote host closed the connection 1170847544 J * lilalinux ~plasma@80.69.41.2 1170848692 J * ZLinux ~ZLinux@88.213.15.157 1170849310 Q * ZLinux Quit: Leaving 1170849329 J * ZLinux ~ZLinux@88.213.15.157 1170849818 J * cdrx ~legoater@cap31-3-82-227-199-249.fbx.proxad.net 1170850651 Q * ema Quit: leaving 1170850813 J * kadeline ~orange@cognet.ci0.org 1170850827 M * kadeline hi! 1170850954 M * kadeline i have a physical computer under fedora core 6 and a vserver running under fedora core 5. 1170850972 M * kadeline when i do a vrpm or a vyum on the vserver, it hangs. 1170850983 M * kadeline and i can't find why. 1170850995 M * kadeline any idea? 1170851172 M * kadeline vrpm begins to list the packages but never ends 1170851198 M * kadeline i tried to rebuild the db, with no success.. 1170851243 Q * kadeline Quit: Changing server 1170851429 J * kade ~orange@cognet.ci0.org 1170851441 M * kade hi again :) 1170851452 M * kade is there anyone alive? 1170851771 M * kade brb 1170851773 Q * kade Quit: leaving 1170851946 J * kadeline ~agilles1@195.83.225.167 1170851952 M * kadeline hey 1170852251 M * kadeline someone? 1170852465 M * eyck someone is not here at the moment 1170852481 M * kadeline lol 1170852557 Q * kadeline Remote host closed the connection 1170852669 J * kadeline ~agilles1@195.83.225.167 1170852675 M * kadeline erf ;) 1170852701 M * kadeline eyck: do you have any id for resolving my vrpm trouble? 1170852931 M * eyck try strace, look where it's stopping 1170853075 M * kadeline that s strange bacause it begins to print the package list on the screen and hangs in the middle 1170853204 Q * kadeline Remote host closed the connection 1170853436 J * kadeline ~agilles1@195.83.225.167 1170853437 M * kadeline err 1170853969 J * FleStA ~besen@lan6-136-138.interbild.net 1170853971 P * FleStA 1170854931 J * marcfiu ~mef@aegis.CS.Princeton.EDU 1170854950 M * daniel_hozac kadeline: so rm -f /vservers/.pkg//rpm/state/__db*; vrpm -- --rebuilddb doesn't do the trick? 1170854956 M * daniel_hozac hey marcfiu 1170855052 J * Bloodman ~blood@66.172.213.129 1170855055 M * kadeline nope 1170855062 M * kadeline unfortunately 1170855068 Q * Bloodman 1170855070 M * daniel_hozac kadeline: note that the rm -f is important. 1170855172 M * kadeline i got some error message 1170855219 M * kadeline the temporary db /vservers/guest/.rpmdb already exists 1170855270 M * kadeline and when i try to remove it, i got another error : already in use. 1170855293 M * daniel_hozac seems something is rather messed up on your system. 1170855320 M * kadeline indeed 1170855323 M * kadeline er 1170855336 M * kadeline what do you suggest? 1170855487 M * daniel_hozac does vps faux show any rpm processes? 1170855524 M * kadeline yes it does 1170855574 M * baldy_ Bertl_zZ: aaround... 1170855578 M * baldy_ dont seems sohehe 1170855590 M * daniel_hozac kadeline: try killing them 1170855593 M * kadeline hi baldy 1170855594 M * kadeline ok 1170855620 M * baldy_ huhu kadeline 1170855624 N * baldy_ baldy 1170855670 M * kadeline daniel_hozac: done 1170855709 M * daniel_hozac kadeline: and still not working? 1170855791 M * kadeline nope, .rpmdb seems still busy 1170855829 M * daniel_hozac and the rpm processes did die? 1170855835 M * kadeline yep 1170855841 M * daniel_hozac that's odd. 1170855901 M * daniel_hozac Bertl_zZ: looks like there were multiple issues with the legacy stuff. the second vc_new_s_context now fails, due to the CAP_SYS_ADMIN check in the switch. 1170855917 M * kadeline yey! 1170855950 M * kadeline found some yum process that was running 1170856136 M * kadeline but i still cant rebuild the db 1170856472 M * kadeline if i remove /vserver/guest/.rpmdb i got a message telling: secure-mount: chdir("/.rpmdb"): No such file or directory 1170856487 M * kadeline if i create it again it tells me : 1170856489 M * daniel_hozac yes. you shouldn't remove it. 1170856550 M * kadeline it tells that .rpmdb already exists 1170856561 M * kadeline but the directory is empty.. 1170856595 M * renihs ls -la 1170856616 M * Radiance which cap should i give a guest to manipulate the mtu within a guest ? 1170856617 M * daniel_hozac and if your root directory is empty, you have much more serious issues. 1170856627 M * daniel_hozac Radiance: you simply shouldn't. 1170856635 M * daniel_hozac networking happens on the host. 1170856639 M * Radiance ewww :) 1170856664 M * Radiance so can i change the mtu for a guest interface from the host ? 1170856679 M * daniel_hozac there's no such thing as a guest interface. 1170856683 M * Radiance like gre0 1170856722 M * Radiance yeah 1170856766 M * kadeline daniel_hozac: .rpmdb should contain the same thing as /vserver/.pkg/guest/rpm/state ? 1170856775 M * daniel_hozac only in the rpm/yum namespace. 1170856910 M * kadeline sorry but i m not sure i understand 1170857131 M * daniel_hozac Bertl_zZ: delta-legacy-fix01 + http://people.linux-vserver.org/~dhozac/p/k/delta-legacy-fix02.diff seems to be enough. 1170857148 M * daniel_hozac kadeline: vserver uses namespaces to separate the mounts of one guest from another. 1170857148 J * ema ~ema@lart.galliera.it 1170857169 M * kadeline yep 1170857172 M * daniel_hozac kadeline: the namespaces are also used for vyum/vrpm, to mount the .rpmdb only in their namespaces. 1170857179 M * kadeline ok 1170857510 M * kadeline thx for your help anyway 1170857529 M * kadeline i think i m going to reinstall it from scratch 1170857537 M * kadeline :) 1170857541 M * kadeline got to go now. 1170857545 M * kadeline bye 1170857548 P * kadeline 1170858859 J * hiaslboy ~hiaslboy@85.127.227.37 1170859702 J * chand ~chand@m167.net81-64-156.noos.fr 1170859705 P * chand 1170859741 J * chand ~chand@m167.net81-64-156.noos.fr 1170860569 Q * hiaslboy Quit: Kopete 0.10.2 : http://kopete.kde.org 1170861175 J * vertyhe ~ron@ner-as27993.alshamil.net.ae 1170861179 M * vertyhe vmware issues? 1170861183 Q * TrueBrain Ping timeout: 480 seconds 1170861188 M * daniel_hozac huh? 1170861212 M * vertyhe just looking for a vmware channel 1170861227 M * daniel_hozac this is not it. 1170861231 P * vertyhe 1170861237 M * nou lol 1170861254 M * nou we could have told him to go read news.vmware.coł 1170861261 M * nou m 1170861677 M * baldy lol 1170862064 Q * DavidS Quit: Leaving. 1170862166 A * cehteh is playing with kvm ... actually without installing windows isnt yet supported, lets see if running the imagge will work then 1170862949 Q * cdrx Remote host closed the connection 1170863077 J * cdrx ~legoater@cap31-3-82-227-199-249.fbx.proxad.net 1170863149 Q * cdrx 1170863162 J * cdrx ~legoater@cap31-3-82-227-199-249.fbx.proxad.net 1170863338 M * baldy Bertl_zZ: sleeps too much hehe 1170863529 M * daniel_hozac how so? 1170863676 M * baldy i see him since this moring as zZ marked hehe 1170863709 M * baldy we spoke over some donates... 1170863721 A * baldy will sponsor somethink 1170863727 M * daniel_hozac well, yeah. but 13 hours of sleep isn't too much... 1170863735 M * baldy lol? 1170863738 M * baldy 6 hours is oke ;) 1170863753 M * Hollow 13 is definitely too much :) 1170863760 M * Hollow morning btw 1170863760 M * daniel_hozac (and i assume he hasn't actually been asleep for a few hours) 1170863764 M * daniel_hozac morning Hollow. 1170863769 Q * mrp- Quit: What's a script? 1170863775 M * daniel_hozac 15 hours of sleep is about right :) 1170863788 M * baldy strange... 1170863797 M * baldy 15h of sleep... 1170863811 M * Hollow daniel_hozac: i spent the whole day adding vserver support to valgrind, interested in testing it? 1170863819 M * daniel_hozac hmm? 1170863828 M * Hollow you know valgrind? 1170863834 M * daniel_hozac (i've never used valgrind before, it's some sort of debugger, right?) 1170863850 M * Hollow memory debugger, call graphs, profiling .. 1170863856 M * daniel_hozac right... 1170863867 M * daniel_hozac so what does the vserver support do? 1170863882 M * daniel_hozac follow processes into the contexts? 1170863883 M * Hollow add the sys_vserver wrapper so valgrind can handle our syscalls 1170863891 M * daniel_hozac ah, okay. 1170863924 M * Hollow http://home.xnull.de/work/gentoo/cvs/dev-util/valgrind/files/valgrind-3.2.3-sys_vserver.patch 1170864070 M * daniel_hozac was this automated in any way? because if not, i'm sorry. must've taken your hours to add the header stuff... 1170864089 M * Hollow well, it was automated by sed and vim :) 1170864094 M * daniel_hozac okay ;) 1170864160 M * Hollow daniel_hozac: just wanted to conform the previous valgrind declaration of kernel headers, but it seems to work quite well here, valgrind know accept the vserver syscall :) 1170864209 M * daniel_hozac sounds cool! 1170864255 M * Hollow will submit it upstream asap 1170864302 M * daniel_hozac how does valgrind work? strace? 1170864398 M * cehteh virtual machine 1170864517 M * Hollow i'm not sure how valgrind does work, but at least for the syscalls it seems to be a simple wrapper that intercepts syscalls, performs some valgrind sepcific pre/post functions and calls the syscall unmodified afterwards .. 1170864545 M * cehteh valgrind is a virtual machine i meant 1170864570 M * Hollow yeah, didn't look very detailed at the valgrind internas, but virtual machine mackes sense :) 1170864876 M * daniel_hozac hmm, cool. 1170865076 M * cehteh needs to intrumentate every cpu instuction .. hence its so damn slow 1170865562 M * matti Hi folks! 1170865590 M * meandtheshel1 Hollow: that's cool - are you going to maintain the vserver-valgrind patch from now on? the location (URL above) will remain the same? 1170865648 M * Hollow meandtheshel1: well, i can update it from time to time.. but won't maintain a vserver patch for valgrind in gentoo 1170865657 M * Hollow the plan is to support it upstream 1170865661 M * Hollow :) 1170865669 M * matti Yeah! 1170865678 J * stefani ~stefani@tsipoor.banerian.org 1170865695 A * matti want support for Gentoo Coffee Making Subroutine in upstream ;] 1170865695 M * meandtheshel1 good plan - so put it in the ftp area :) 1170865730 M * matti http://it.slashdot.org/article.pl?sid=07/02/07/0146245 1170866582 Q * cdrx Quit: Leaving 1170866601 J * cdrx ~legoater@cap31-3-82-227-199-249.fbx.proxad.net 1170866753 Q * stefani synthon.oftc.net strange.oftc.net 1170867193 M * meandtheshel1 matti: folks should use http://en.wikipedia.org/wiki/OpenWengo - a really terrific thing - best about it is its FireFox extension - means also my mum can wengo_with_me :) 1170867264 J * stefani ~stefani@flute.radonc.washington.edu 1170867409 Q * cdrx Quit: Leaving 1170867468 J * cdrx ~legoater@cap31-3-82-227-199-249.fbx.proxad.net 1170867756 J * badari ~badari@bi01p1.co.us.ibm.com 1170867808 N * DoberMann DoberMann[PullA] 1170867812 M * cdrx daniel_hozac, ping 1170867826 M * daniel_hozac pong 1170867832 M * cdrx hi 1170867837 M * daniel_hozac hello 1170867883 M * cdrx could you remind me what where the issues with 2.6.19 when you were mounting a NFS share after having unshared the mnt namespace ? 1170867907 M * cdrx without the patch we added that really broke things 1170867917 M * daniel_hozac NFS tries to lock current->sighand. 1170867928 M * cdrx yep 1170867944 M * daniel_hozac and it sleeps before that, which gives the parent ample opportunity to reap the process. 1170867950 M * cdrx even without our patch ? 1170867963 M * daniel_hozac which patch are we talking about? 1170868003 M * cdrx the one that moved exit_task_namespace() below exit_notify() 1170868031 M * daniel_hozac that was done for 2.6.19, no? 1170868048 M * daniel_hozac with exit_task_namespaces before exit_notify, there's no problem. 1170868087 M * cdrx ok. that's what i wanted to know. I thought you mentionned some issues with earlier kernels 1170868100 M * daniel_hozac no. 2.6.18 is unaffected, just 2.6.19+. 1170868101 M * cdrx thanks ! 1170868106 M * daniel_hozac np. 1170868129 J * bonbons ~bonbons@83.222.37.103 1170868139 M * cdrx we're looking for a clean fix 1170868229 Q * stefani synthon.oftc.net xenon.oftc.net 1170868229 Q * Johnsie synthon.oftc.net xenon.oftc.net 1170868229 Q * DreamerC synthon.oftc.net xenon.oftc.net 1170868229 Q * nebuchadnezzar synthon.oftc.net xenon.oftc.net 1170868242 J * hallyn ~xa@cpe-72-179-43-119.austin.res.rr.com 1170868260 M * daniel_hozac moving the pid namespace out of the nsproxy seemed like the cleanest one to me. 1170868269 M * daniel_hozac it needs to live longer than the others. 1170868300 M * cdrx yes 1170868318 M * cdrx we had a request from the top penguins to add nsproxy 1170868338 M * cdrx which makes the task_struct life easier in some ways 1170868345 M * cdrx but not always ... 1170868357 M * daniel_hozac yeah, the overhead might not be worth it... 1170868426 M * cdrx agree, but it does make the unshare easier and keeps the task_struct smaller 1170868439 M * hallyn and typical clone quicker 1170868450 M * daniel_hozac well, i meant the overhead of moving the pid_ns out of it ;) 1170868455 M * hallyn ah 1170868532 M * hallyn well we may do that short term, use accessor functions to get task->nsproxy->pid_ns or task->pid_ns, and then if we find a solution at the nfs server and can move the pidns back in, keep a simple localized change at pid_namespace.h/c 1170868538 J * DreamerC ~dreamerc@125-225-98-13.dynamic.hinet.net 1170868571 M * daniel_hozac is everything being moved over to kthreads? 1170868582 M * daniel_hozac i guess that might take care of the NFS server stuff. 1170868604 M * cdrx hmm, not sure 1170868611 M * daniel_hozac (as it'd require someone poking at that code anyway) 1170868629 M * cdrx but yes not a lot of kthread are left (nfs is resisting though) 1170868631 M * hallyn that'd be nice 1170868690 M * daniel_hozac did anyone figure out what NFS is doing to current's signals anyway? 1170868726 J * Johnsie ~jdlewis@jdlewis.org 1170868726 J * nebuchadnezzar ~nebu@zion.asgardr.info 1170868743 Q * SNy Remote host closed the connection 1170868769 J * stefani ~stefani@flute.radonc.washington.edu 1170869050 J * SNy a6de27b757@bmx-chemnitz.de 1170869131 Q * DreamerC Quit: leaving 1170869147 J * DreamerC ~dreamerc@125-225-98-13.dynamic.hinet.net 1170869377 N * Bertl_zZ Bertl 1170869381 M * Bertl morning folks! 1170869418 M * daniel_hozac morning Bertl! 1170869430 M * cdrx morning 1170869447 M * Bertl cdrx: so nsproxy is becoming part of the task struct? 1170869506 M * cdrx what do you mean ? nsproxy is part of the task struct. 1170869544 M * Bertl yeah, but I mean instead of nsproxy *p, nsproxy p :) 1170869548 M * cdrx daniel_hozac, NFS uses kernel_thread task and signals. if it used kthread and no signals it would make our life easier. 1170869552 M * cdrx ah 1170869556 M * cdrx :) 1170869570 M * cdrx well, we're brain storming on that topic 1170869615 A * cdrx amazed with Bertl timezone 1170869634 J * hiaslboy ~hiaslboy@85.127.227.37 1170869664 M * hiaslboy Hallo Bertl :-) hello everybody :-) 1170869673 M * Bertl hey hiaslboy! 1170869691 M * Bertl cdrx: yeah, unfortunately I missed the 'shopping time' today :/ 1170869705 M * cdrx no food for breakfast ? 1170869710 M * Bertl cdrx: i.e. shops are already closed here ... 1170869716 M * daniel_hozac haha 1170869722 M * Bertl I have enough in the deep freezer ... 1170869736 M * cdrx well it's beer time here in france 1170869754 M * hiaslboy Bertl: bad luck ... here the shops are open till 7:30 (at least some :-) 1170869764 A * Bertl is getting some frozen mammoth from the basement ... 1170869928 M * hallyn yum 1170870415 Q * dna Read error: Connection reset by peer 1170870563 J * dna_ ~naucki@153-227-dsl.kielnet.net 1170870632 M * Bertl cdrx: so what will be left in nsproxy? 1170870635 J * dna ~naucki@153-227-dsl.kielnet.net 1170870635 Q * dna Read error: Connection reset by peer 1170870642 M * Bertl uts and ipc? 1170870670 M * cdrx yes and futur pid and net 1170870731 M * cdrx no decision taken yet but we might as well start a thread on the topic 1170870739 A * cdrx gotta go 1170870750 M * cdrx see ya 1170870754 M * Bertl cya 1170870925 N * DoberMann[PullA] DoberMann 1170870950 Q * ntrs_ Remote host closed the connection 1170871179 N * debugger rgl 1170871204 Q * thunder1 Ping timeout: 480 seconds 1170871214 J * thunder1 ~thu@anonymisierungsdienst2.foebud.org 1170871347 M * Bertl daniel_hozac: btw, I reverted the nfs fix from the 2.6.20 patch, as I assume the issue was fixed there ... do you know if that is true? 1170871377 M * Bertl daniel_hozac: another thing, how did the legacy workaround work out? 1170871396 J * tso ~tso@238-048.adsl.pool.ew.hu 1170871398 M * tso hi all 1170871408 M * Bertl welcome tso! 1170871414 M * tso hi Bertl 1170871420 M * tso i got some oops 1170871425 M * tso ;) 1170871429 M * Bertl what kernel? 1170871440 M * tso i'll try to find that is it grsec or vserver related 1170871466 M * tso 2.6.19.2 with grsec and vserver 1170871468 M * tso patch-2.6.19.2-vs2.2.0-rc10-grsec2.1.10-20070203.diff 1170871473 M * Bertl check with 2.6.19.3-vs2.2.0-rc11 1170871493 M * daniel_hozac Bertl: http://people.linux-vserver.org/~dhozac/p/k/delta-legacy-fix02.diff was needed as well. 1170871536 M * daniel_hozac Bertl: yeah, looks like it should be fixed. 1170871549 M * tso Bertl, ok... 1170871549 M * Bertl 0? 1170871570 M * daniel_hozac i was unsure what the values meant, it needs to be 0 or 1 though. 1170871586 M * daniel_hozac otherwise the CAP_SYS_ADMIN check disallows it. 1170871587 M * Bertl daniel_hozac: that should be a security issue 1170871610 M * Bertl i.e. even normal users can execute it :) 1170871620 M * daniel_hozac the capable(CAP_CONTEXT) should prevent that, no? 1170871695 M * Bertl will check that after dinner ... 1170873366 Q * chand Quit: chand 1170873682 M * cryptronic hi all 1170873862 M * daniel_hozac hello 1170873933 M * cryptronic daniel_hozac, maybe you could help me ;) do you know a good ressource where the Linux Resource Limits are explained? 1170874053 M * cryptronic because i do not understand all of these limits explained here: http://linux-vserver.org/Resource_Limits 1170874319 M * daniel_hozac which ones? 1170874438 M * cryptronic SIGPENDING, MSGQUEUE FSIZE STACK CORE 1170874521 M * daniel_hozac SIGPENDING limits the number of signals you can queue for a process. 1170874594 M * daniel_hozac MSGQUEUE limits the number of bytes a process can put in an IPC message queue. 1170874679 M * daniel_hozac FSIZE limits the size of files, as it says. 1170874700 M * daniel_hozac STACK limits the stack size of the processes. 1170874719 M * daniel_hozac CORE basically lets you choose how big core dumps you want. 1170874737 M * daniel_hozac (if it's over that size, the core dumping application will just terminate) 1170874763 M * cryptronic fsize - does this limit fs files or "files" for cpu 1170874780 M * daniel_hozac what? 1170874823 M * cryptronic when is it useful to set fsize 1170874844 M * daniel_hozac i have no idea. 1170874861 M * cryptronic ah ok ;) 1170874867 M * daniel_hozac i can't think of any use case for it... 1170874873 M * Bertl when you are doing batch jobs and have no disk limits :) 1170874890 M * Bertl similar to the cpu limit, it's ancient 1170874929 M * cryptronic ah ok, because from openvcp we decide which limit's are useful for setting them via webinterface 1170874983 M * daniel_hozac supporting them all, in case someone needs them, makes the most sense to me. 1170874993 M * daniel_hozac (that's what util-vserver does) 1170875070 M * daniel_hozac maybe hide them in some sort of advanced tab or whatever. 1170875085 M * cryptronic i think we make rss, nproc, nofile, memlock, as, locks default visible and if you want to set them all you can make the rest visible 1170875172 M * cryptronic daniel_hozac, where can i see the default values for these rlimits or doesn' exist default values? 1170875180 M * daniel_hozac there are no defaults. 1170875212 M * Bertl default is inf (unlimited) 1170875226 M * cryptronic ah thanks 1170875514 M * cryptronic with vlimit i can change rlimit's on the fly without have to restart a guest right? 1170875529 M * daniel_hozac yep 1170875538 M * Bertl the per guest rlimits yes, the ulimits, not 1170875606 M * cryptronic ok thanks 1170876407 J * xp_prg ~xp_prg@ftp.microvu.com 1170876411 M * xp_prg wow hi all! 1170876434 M * xp_prg I setup a vserver with centos4 on debian, it does not have rpmbuild, anyone know how I get that? 1170876471 M * Bertl the guest? 1170876474 M * xp_prg yes 1170876486 M * Bertl you installed it via network? 1170876515 M * xp_prg yes 1170876528 M * Bertl so you have rpm and friends installed on the host 1170876541 M * Bertl and can use vrpm and friends to install more 1170876563 M * xp_prg server centos4 build -m yum --context (8xxx) --hostname (name) --interface pricelist=eth1:172.168.1.(x)/24 -- -d centos4 1170876580 M * xp_prg does vrpm do an rpmbuild of a source rpm? 1170876584 J * TrueBrain truelight@openttd.org 1170876585 M * Bertl so, just use vyum 1170876607 M * Bertl something like vyum install 1170876609 M * xp_prg but this is a source rpm 1170876617 M * xp_prg I use rpmbuild on centos proper 1170876630 M * Bertl so install rpmbuild in the guest :) 1170876639 M * xp_prg Bert1 I am trying to do that 1170876643 M * xp_prg how might I do that? 1170876651 M * Bertl as I said, with vyum :) 1170876680 M * xp_prg I have yum installed inside of my centos4 guest 1170876688 M * xp_prg I did a yum search rpmbuild and nothing came up :( 1170876701 M * Bertl well, that's a centos issue, no? 1170876718 M * Bertl you have to ask centos folks how the package is called 1170876720 M * xp_prg ya I guess, I will try this them 1170876723 M * xp_prg thanks man :) 1170876727 M * Bertl np 1170876745 M * xp_prg Bert1 have you encountered a problem where make on a debian etch guest said that it could not create a file? 1170876755 M * xp_prg running as root? 1170876781 M * Bertl can happen, even on normal linux boxes 1170876796 M * Bertl e.g. an immutable file cannot be overwritten 1170876806 M * xp_prg how can I fix that? 1170876828 M * Bertl remove the immutable flag 1170876838 M * Bertl but there could be other reasons 1170876838 M * xp_prg so chmod 777 won't do it? 1170876844 M * Bertl nope 1170876849 M * xp_prg what is the command? 1170876874 M * Bertl chattr -i 1170876892 M * xp_prg hmm... ok 1170876912 M * daniel_hozac the package is called rpm-build. 1170877144 M * baldy ahh Bertl ;) 1170877160 M * Bertl hey baldy! 1170877186 M * baldy did u need a ded. server? i can sponsor the projekt a xeon server if u wish 1170877214 M * Bertl sounds good, I guess we can make use of that, just two 'minor' requirements 1170877222 Q * xp_prg Quit: Ciao! 1170877242 M * Bertl we need some way to get a serial console there (for debugging) 1170877256 M * Bertl we probably need a way to reset it remotely 1170877260 J * xp_prg2 ~xp_prg2@ftp.microvu.com 1170877272 M * baldy no prob :) 1170877290 M * Bertl excellent! 1170877305 M * baldy server is free on friday for u 1170877324 M * Bertl great! 1170877334 M * baldy what about ips? 1170877339 M * baldy how many ips are needed? 1170877360 M * Bertl I guess 2-3 would be enough for testing purposes 1170877372 M * baldy oki 1170877387 M * Bertl would be nice to get some in two different networks if that is an option? 1170877395 M * baldy jop 1170877435 M * Bertl and let me know all the details so that we can add you (folks?) to the Hall'o'Fame 1170877456 M * baldy jops hehe 1170877481 M * baldy i can register zwo /29 for u 1170877484 M * baldy if u wish 1170877492 M * baldy on the vserver projects name 1170877494 M * Bertl excellent ... 1170877586 M * baldy patch for 2.6.20 out? 1170877598 M * Bertl yep, but only for testing 1170877600 M * daniel_hozac http://vserver.13thfloor.at/Experimental/patch-2.6.20-vs2.2.0-pre1.diff 1170877605 M * Bertl tx 1170877619 M * nebuchadnezzar thanks daniel_hozac 1170877625 M * nebuchadnezzar I'll try that ASAP 1170877634 M * daniel_hozac well, i just pasted the URL ;) 1170877663 M * nebuchadnezzar yes, thanks for that :-) 1170877712 M * nebuchadnezzar arf, I'm waiting the 2.3.0.x one :-) 1170877761 M * Bertl you can probably do a delta between the branches on 2.6.19 and use that on 2.6.20 1170877947 M * baldy whats the best Timer frequency choice for the vServer? 1170877985 M * Bertl depends on the usage pattern of the guests 1170878019 M * Bertl but usually a lower Hz value means less context switches 1170878048 M * Bertl so, unless you really need high responsiveness 1170878053 M * Bertl you would go for 100Hz 1170878100 M * baldy TCP: MD5 Signature Option support 1170878103 M * baldy thats cool 1170878146 M * matti baldy: :))) 1170878155 M * matti Ops. 1170878156 M * matti Bertl: :) 1170878163 M * matti ;p 1170878163 M * Bertl hey matti! 1170878206 M * matti Hello! 1170878208 M * matti ;D 1170878502 M * daniel_hozac anyone know off the top of their head how to do an internal link with a non-link text? 1170878545 M * daniel_hozac ... on the wiki 1170878690 J * FireEgl Proteus@68.220.222.136 1170878712 Q * tso Quit: BitchX-1.1-final -- just do it. 1170878728 J * donuto 97c873a216@83.149.112.45 1170878754 M * Bertl welcome donuto! 1170878784 M * donuto hi Bertl 1170878829 M * donuto i think i am forced to use some capabilety since openvpn can not set tx/mtu saying operatione not permitted :( 1170878880 M * Bertl hmm? 1170878886 M * daniel_hozac why is it trying to? 1170878904 M * baldy Bertl: is it possible to use pptpd inside a vps? 1170878912 M * Bertl daniel_hozac: besides that, on the host this might only be related to your network card driver? 1170878918 M * baldy pptpd opens pppX network device 1170878925 M * Bertl s/daniel_hozac/donuto 1170878944 M * donuto i get tis message: Note: Cannot set tx queue length on tunVPN0: Operation not permitted (errno=1) and /sbin/ip link set dev tunVPN0 up mtu 1500 SIOCSIFMTU: Operation not permitted 1170878959 M * daniel_hozac so you should disable that. 1170878969 M * daniel_hozac rather than give your guest the ability to do whatever it wants with the network stack. 1170878973 M * donuto this when i try start with openvpn --config server.conf 1170878993 M * Bertl which just means that your server.conf is wrong :) 1170879005 M * baldy daniel_hozac: what is the option for this?! 1170879013 M * donuto i follow examples on openvpn site 1170879024 M * daniel_hozac baldy: option for what? 1170879044 M * baldy your guest the ability to do whatever it wants with the network stack. 1170879046 M * baldy ;) 1170879053 M * baldy so it means 1170879055 M * donuto daniel_hozac, how can i disable this ? 1170879077 M * daniel_hozac donuto: no idea. i haven't used openvpn in a guest. 1170879080 M * donuto without using (not sure) cap_net_admin or raw ? 1170879094 M * daniel_hozac baldy: CAP_NET_ADMIN as per http://linux-vserver.org/util-vserver:Capabilities_and_Flags 1170879107 M * baldy when i have a device named tunv6 a user is able to do: ip a a ip:/48 dev tunv6? 1170879123 M * donuto i think only root 1170879180 M * baldy user = root in his vps 1170879264 M * baldy Bertl: to u need native v6 on the server? 1170879280 M * baldy s/to/do 1170879297 M * daniel_hozac wouldn't hurt, i guess.. 1170879346 M * xp_prg2 posgres built thanks all! 1170879569 M * baldy Found kernel: /boot/vmlinuz-2.6.20 1170879573 M * baldy drck die duamen 1170879575 M * baldy huchu 1170879595 J * tso ~tso@238-048.adsl.pool.ew.hu 1170879856 M * baldy Linux andrew 2.6.20vs2.2.0-vs2.2.0-pre1 #1 1170879859 M * baldy great ;) 1170879882 M * daniel_hozac Bertl: wow, i can reproduce Jarek's oops now. 1170880022 Q * mnemoc Ping timeout: 480 seconds 1170880086 M * daniel_hozac Bertl: i guess it's because i enabled the CONFIG_DEBUG_RWSEMS and CONFIG_DEBUG_SLAB. 1170880136 M * daniel_hozac Bertl: /proc/virtual/status does not show a decrease in fs_struct's. 1170880512 J * mnemoc ~amery@kilo105.server4you.de 1170880786 M * baldy strange... 1170880804 M * baldy i cant kill any prozess in the vps 1170880818 M * daniel_hozac on 2.6.20? 1170880821 M * baldy yep 1170880833 M * baldy maxom:~# ps aux |grep nagios |wc -l 1170880833 M * baldy 95 1170880833 M * baldy maxom:~# killall -9 nagios 1170880833 M * baldy maxom:~# ps aux |grep nagios |wc -l 1170880835 M * daniel_hozac with vkill, or with kill inside? 1170880835 M * baldy 95 1170880881 M * daniel_hozac use kill and a specific pid. 1170880904 M * baldy maxom:~# kill -9 4945 1170880904 M * baldy maxom:~# ps aux |grep 4945 1170880904 M * baldy sshd 4945 0.0 0.0 0 0 ? Z 21:36 0:00 [sshd] 1170880904 M * baldy root 6371 0.0 0.0 1636 556 pts/0 S+ 21:41 0:00 grep 4945 1170880935 M * PowerKe daniel_hozac: Someone ran into a problem of not being able to run vserver start when /tmp is a link 1170880951 M * daniel_hozac symlinks are bad. 1170880971 M * daniel_hozac they cause all sorts of problems during the vserver ... start process. 1170880986 M * baldy back to 2.6.19 ;) 1170880987 M * daniel_hozac (seen with nice/env on certain distributions) 1170880993 M * PowerKe Didn't cause any problems with util .211 1170881015 M * daniel_hozac that's highly unlikely, as nothing really changed with 0.30.212 in that regard. 1170881111 M * PowerKe error was: vcontext: jailIntoTempDir(): No such file or directory 1170881122 M * PowerKe vcontext: jailIntoTempDir(): No such file or directory 1170881124 M * baldy is there a patch for 2.4.34 kernel? 1170881128 M * PowerKe vcontext: write(): Broken pipe 1170881185 M * baldy i heard some application wont work with 2.6.x kernel 1170881223 M * daniel_hozac PowerKe: yes i know, i read the back log. 1170881230 M * PowerKe ok 1170881245 M * daniel_hozac baldy: talk to eyck. he maintains the 2.4 branch. 1170881264 M * daniel_hozac baldy: but applications that break on 2.6 really need to be fixed.... 1170881275 M * baldy daniel_hozac: yeah i know... 1170881281 M * baldy psyBNC is one of them :( 1170881294 M * daniel_hozac ? why on earth would a bouncer need 2.4? 1170881322 M * baldy daniel_hozac: dont know 1170881327 M * baldy a friend a mine told me 1170881334 M * baldy that psybnc freeze with 2.6.er 1170881356 M * baldy he problem is known in many foroms 1170881360 M * baldy forums 1170881438 M * baldy http://www.serversupportforum.de/forum/irc/9875-psybnc-auf-2-6er-kernel.html 1170881466 Q * donuto Quit: Changing server 1170881516 M * daniel_hozac my german is terrible, but that says it's only when compiled with oident support, no? 1170881612 M * baldy no 1170881637 M * baldy i found in an other forum the the new glib wont work 1170881646 M * baldy 2.3.1 will work with 2.6.er kernel 1170882092 Q * mnemoc Read error: Connection reset by peer 1170882104 J * mnemoc ~amery@kilo105.server4you.de 1170882159 J * ntrs ~ntrs@68-188-55-120.dhcp.stls.mo.charter.com 1170882201 M * ntrs Bertl, what's up with the 2.2.0-rc11 crashing reported on the mailing list? 1170882213 M * daniel_hozac ntrs: i'm trying to track that down now. 1170882217 M * Bertl no idea yet ... 1170882234 M * ntrs I was hoping to upgrade but I guess I'll have to wait. 1170882271 M * Bertl problem is likely to be in all 2.6.19 versions 1170882283 M * Bertl (at least since we fixed the nsproxy leak) 1170882314 M * daniel_hozac Bertl: the current->fs is correct. 1170882350 M * Bertl how did you verify? 1170882395 M * daniel_hozac vxdprintk(VXD_CBIT(misc, 7), "name %s comm %s fs %p", name, current->comm, current->fs); 1170882402 M * daniel_hozac at the beginning of do_path_lookup 1170882405 M * daniel_hozac (before the if) 1170882433 M * Bertl doesn't tell anything .. could be already disposed, no? 1170882436 M * daniel_hozac and a vxdprintk(VXD_CBIT(misc, 7), "old fs %p, new fs %p", fs_new, current->fs); 1170882443 M * daniel_hozac i suppose. 1170882464 M * Bertl you should print the refcount too 1170882830 M * daniel_hozac refcount 1802201963... 1170882839 M * Bertl see 1170882840 M * daniel_hozac (base 10) 1170882849 M * daniel_hozac yeah. 1170883012 M * daniel_hozac ooohhh... 1170883350 M * daniel_hozac http://people.linux-vserver.org/~dhozac/p/k/delta-space-fix03.diff 1170883421 M * Bertl hmm ... 1170883454 M * daniel_hozac otherwise the fs_cur is put twice, while just being get once. 1170883457 M * Bertl that looks wrong ... we should instead get the vxi 1170883499 M * daniel_hozac hmm, get_space... i meant to patch set_space. 1170883594 M * daniel_hozac but somehow, that appears to have fixed the problem?! 1170883621 M * daniel_hozac oh, nevermind me. i meant to fix enter_space. 1170883650 M * Bertl the thing is, you avoid the put when you have the same context, right? 1170883670 M * Bertl s/context/proxy 1170883684 M * daniel_hozac well, the fs_cur is the important change. 1170883695 M * daniel_hozac i just assumed the proxy could have the same problem. 1170883703 M * daniel_hozac (just entering the CLONE_FS space) 1170883718 M * Bertl yes, but what I mean is, you now introduced a case where 1170883725 M * Bertl a) proxy/fs != NULL 1170883736 M * Bertl and b) the put is not done 1170883750 M * daniel_hozac right. 1170883754 M * daniel_hozac because it's done right below. 1170883755 M * Bertl which means that the get/put is unbalanced in this cases 1170883767 M * daniel_hozac no, i'd say it _was_ unbalanced. 1170883781 M * Bertl then you should remove the entire put 1170883790 M * daniel_hozac but we need it. 1170883794 M * Bertl because, if it is unbalanced, it has to be for _all_ cases 1170883797 M * daniel_hozac for the cases where fs_new is not fs_cur. 1170883810 M * Bertl that's a dangerous thought 1170883847 M * Bertl let's assume it is incorrect for the == current case 1170883880 M * Bertl please proof that it _is_ correct for the others, and especially why, as we do get/put in a balanced style 1170883913 M * Bertl but, the real imbalance is in the vxi stuff 1170883945 M * Bertl we hand out the vxi_* to tasks, without grabbing a reference IIRC 1170884011 M * daniel_hozac vxi_*? 1170884025 M * Bertl let's look at enter 1170884038 M * Bertl we do proxy = vxi->vx_nsproxy 1170884049 M * Bertl let's assume that has a refcount of 1 1170884068 M * Bertl let's further assume the task has a different proxy 1170884078 M * Bertl (so your special case does not apply) 1170884144 M * Bertl let's further assume we 'copy' the namespace 1:1 1170884171 M * Bertl i.e. mask=0 1170884192 M * Bertl we then hand out the proxy to the task (with ref=1) 1170884202 M * Bertl and put the existing space 1170884240 M * Bertl now what happens with the next task which enters? 1170884260 M * Bertl it will get the same proxy, with the same refcount of 1 1170884300 M * daniel_hozac is this a theoretical discussion? because AFAICT, we never hand out the entire nsproxy. 1170884313 M * Bertl __vs_merge_nsproxy() 1170884318 M * Bertl return proxy; 1170884333 M * daniel_hozac ah, there we go. 1170884345 M * daniel_hozac yeah, that does look problematic... 1170884361 M * Bertl so, what we actually want is to grab a reference on that 1170884366 N * DoberMann DoberMann[ZZZzzz] 1170884375 M * daniel_hozac indeed. 1170884398 M * Bertl and the current = proxy special case will disappear if the get/put is balanced 1170884418 M * daniel_hozac however, i don't think __vs_merge_nsproxy is ever called with mask = 0. 1170884436 M * Bertl doesn't matter, it's the very same basic principle 1170884445 M * Bertl I just took this one to simplify the case 1170884473 M * Bertl nevertheless, we could do a check upfront 1170884484 M * Bertl and avoid replacing one with the same completely 1170884501 M * Bertl but I think that will really complicate the procedures 1170884547 M * Bertl so the simpler way is to get the get/put balanced 1170884554 M * daniel_hozac right. 1170884560 M * Bertl which will lead to a 'get' at the beginning, and the 1170884569 M * Bertl put will work in the a == b case too 1170884593 M * daniel_hozac hmm, not a get in __vs_merge_nsproxy? 1170884610 M * daniel_hozac the problem is that __vs_merge_nsproxy doesn't always get the result. 1170884614 M * daniel_hozac is it not? 1170884631 M * Bertl the merge is one place where it is wrong 1170884634 M * daniel_hozac (and same is true for the fs_struct. 1170884637 M * Bertl we have a line there: 1170884642 M * Bertl * will hold a reference on the result. 1170884653 M * Bertl this is _very_ important 1170884668 M * daniel_hozac yeah. 1170884670 P * DEac- 1170884809 Q * nou Remote host closed the connection 1170884884 J * Aiken ~james@ppp220-70.lns2.bne1.internode.on.net 1170884948 M * Bertl welcome Aiken! 1170884980 J * Piet_ hiddenserv@tor.noreply.org 1170885070 M * tso Bertl 2.2.0-rc11 seems to be ok for me, but meantime i have found that my oopses related to grsec, they will release a new version where this problems will be fixed 1170885095 M * Bertl tso: okay, tx for the feedback 1170885109 M * tso Bertl thx for your great work ;) 1170885186 M * Bertl it's not just me as you see ... 1170885209 M * tso yes of course it's not 'just' for you... 1170885277 M * daniel_hozac Bertl: so, does http://people.linux-vserver.org/~dhozac/p/k/delta-space-fix04.diff look better? 1170885284 M * daniel_hozac it seems to fix the problem. 1170885317 M * Bertl yep, much better 1170885332 M * Bertl and we still need the get for the proxy 1170885346 M * Bertl (even if mask=0 is not exercised yet) 1170885359 Q * Piet Ping timeout: 480 seconds 1170885381 M * daniel_hozac right. 1170885539 Q * hiaslboy Ping timeout: 480 seconds 1170885724 M * daniel_hozac http://people.linux-vserver.org/~dhozac/p/k/delta-space-fix05.diff 1170885749 M * Bertl great! tx! 1170885766 M * Bertl btw, do we know if the nfs stuff is fixed in 2.6.20? 1170885774 Q * ema Quit: leaving 1170885814 M * daniel_hozac know? i haven't gotten around to testing 2.6.20 just yet, but AFAICT it reverts to the 2.6.18 sequence in do_exit. 1170885831 M * Bertl ah, so that should be fine then, yes? 1170885840 M * daniel_hozac yep. 1170885857 M * Bertl great, will integrate your fixes and do an rc12 1170885882 M * daniel_hozac great! 1170886351 Q * bonbons Quit: Leaving 1170886583 J * mire ~mire@10-167-222-85.adsl.verat.net 1170887271 J * gerrit ~gerrit@c-67-160-146-170.hsd1.or.comcast.net 1170887428 Q * marcfiu Quit: Download Gaim: http://gaim.sourceforge.net/ 1170887516 T * Bertl 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 ;) 1170887521 M * Bertl *2.2.0-rc12 1170887539 M * Bertl (includes a modified version of the legacy fix :) 1170887875 N * mine Medivh 1170887933 Q * lilalinux Remote host closed the connection 1170888315 P * Hollow Konversation terminated! 1170888324 J * Hollow ~hollow@styx.xnull.de 1170888331 M * Hollow Bertl: http://paste.linux-vserver.org/1109 1170888391 M * Hollow ah.. now i see the lookup forbidden blah .. 1170888397 M * Hollow nevertheless theres the BUG 1170888481 M * daniel_hozac free'ing with at least one task... 1170888555 M * Bertl Hollow: that's the complete run? 1170888571 M * Hollow yep, the first 5 lines or so don't count 1170888594 M * Bertl x_clear_persistent 2x 1170888607 M * Bertl *vx_clear_persistent even 1170888635 M * Hollow yeah, somehow the shutdown helper even runs before startup has finished 1170888638 M * Bertl Hollow: are you changing the persistent stace from within? 1170888643 M * Hollow no 1170888647 M * Bertl i.e. from within the helper? 1170888676 M * Hollow the helper clears persistent at exit 1170888682 M * Bertl ah, that's it 1170888700 M * Bertl we need a check that it is still set before clearing it on helper failure 1170888704 M * Hollow well, i can move that out of the helper 1170888718 M * Bertl nope, that has to be handled at kernel level 1170888722 M * Hollow ok 1170888732 M * Bertl (otherwise we run into that again sooner or later) 1170888742 Q * dna_ Quit: Verlassend 1170888767 M * Bertl Hollow: but you could verify that this is the cause by _not_ doing that for a test 1170888769 M * Hollow basically it goes like this: vcd creates context with persistent, helper calls vcd to do the setup, then starts init and clears persistent 1170888781 M * Hollow not doing what? 1170888809 M * Bertl clearing the persistent state 1170888833 M * Hollow ok, will do 1170889040 M * Hollow Bertl: yep, if i don't clear persistent in the helper it works (i.e. if the helper fails the persistent is cleared once) 1170889099 M * Hollow btw, am i supposed to use dlimits + cow at the same time? 1170889111 J * debugger ~Rui@84.90.10.107 1170889111 M * daniel_hozac why not? 1170889159 M * Hollow this helper bug was triggered by a previous tagged guest .. all newly created templates had the same xid .. 1170889180 Q * attila_ Read error: No route to host 1170889216 M * Hollow this made the helper fail: [52307.775685] vxW: xid=2 did lookup hidden ffff8101378f7aa8[#3,48542] /sbin/init. 1170889221 M * daniel_hozac probably you changed too many xids then ;) 1170889238 M * daniel_hozac you should only tag the files that belong to the guest. 1170889245 M * Hollow but if i tag a linked copy all copies will be changed or unlinked, no? 1170889253 M * Hollow ah, hm .. 1170889257 M * daniel_hozac all copies will be changed. 1170889265 M * daniel_hozac we don't unlink on vc_set_iattr. 1170889267 M * Hollow so the unified files do not count to the limits .. 1170889272 M * daniel_hozac nope. 1170889283 M * daniel_hozac once you break the link, they will. 1170889288 M * Hollow yeah 1170889368 M * daniel_hozac i recently added an option to chxid to only change the xid of non-unified files. 1170889390 M * Hollow yeah, i probably should add this as well (and probably make it the default) 1170889475 M * litage morning guys! 1170889498 M * Hollow ah, cow is so awesome .. http://paste.linux-vserver.org/1110 :) 1170889499 Q * rgl Ping timeout: 480 seconds 1170889506 M * litage so, my vserver guests don't have any 'context' files, but they've been running fine for 6+ months. all of a sudden though, one of them fell over and refuses to bind its ip address 1170889512 M * daniel_hozac hehe, nice! 1170889525 M * Hollow (warm cache though) 1170889539 M * daniel_hozac yeah, i figured. 1170889554 M * daniel_hozac just traversing a guest's tree usually takes longer than a second ;) 1170889584 M * litage i've restarted the guest, and rebooted the host, but the guest still fails to bind its ip address. what can i do? 1170889600 M * daniel_hozac litage: just one guest? 1170889606 M * litage daniel_hozac: yep, just one 1170889609 M * litage the rest are running fine 1170889612 M * Hollow daniel_hozac: on cold cache it typically took 3-4 seconds.. (intel quad-core that is) 1170889627 M * daniel_hozac hehe, damn. 1170889643 M * daniel_hozac litage: and what makes that one so special? 1170889657 M * Hollow so, compared to like 5 minutes tarball unpacking .. ;) 1170889667 M * litage daniel_hozac: no idea. in fact, it hadn't been used since the beginning of december 1170889702 M * litage daniel_hozac: but 2 days ago, i noticed i couldn't ssh into it. so i investigated and found sshd wasn't running [because it couldn't bind an ip address] 1170889727 M * daniel_hozac so what does tail /etc/vservers/*/interfaces/*/ip say? 1170889732 M * daniel_hozac (paste.linux-vserver.org) 1170889793 M * litage daniel_hozac: 'golive' is the problem guest; http://rafb.net/p/k2pEUe96.html 1170889863 M * daniel_hozac no other interfaces? 1170889899 M * litage daniel_hozac: nope. just the one 1170889912 M * daniel_hozac and no other guest or the host is binding on that address? 1170889970 M * litage daniel_hozac: correct: http://rafb.net/p/5XH6k347 1170890021 M * xp_prg2 where should I paste? 1170890028 M * daniel_hozac you notice how the address isn't in that paste, right? 1170890040 M * daniel_hozac xp_prg2: paste.linux-vserver.org 1170890060 M * litage daniel_hozac: yes 1170890073 M * xp_prg2 http://paste.linux-vserver.org/111 1170890078 M * litage daniel_hozac: doesn't that mean that the ip address for the guest is not bound? 1170890080 M * xp_prg2 vyum problem I need help with please 1170890080 M * daniel_hozac litage: i assume manually adding the address makes it work? 1170890102 P * stefani I'm Parting (the water) 1170890108 M * daniel_hozac xp_prg2: did you assign your guest an IP address and configure /etc/resolv.conf in the guest? 1170890171 M * litage daniel_hozac: haven't tried manually adding the address as i don't know how to 1170890187 M * daniel_hozac ip addr add
/ dev ethX 1170890227 M * xp_prg2 let me check hold on please :) 1170890247 Q * Aiken Quit: Leaving 1170890271 M * xp_prg2 gave it the wrong ip address oops sorry thanks 1170890433 M * Hollow Bertl, daniel_hozac: http://bugs.kde.org/show_bug.cgi?id=141366 1170890486 M * daniel_hozac great! 1170890532 M * xp_prg2 http://paste.linux-vserver.org/1112 1170890541 M * xp_prg2 how do I get rid of this extra ip adress I don't want? 1170890614 M * daniel_hozac ip addr del
/, should suffice. 1170890708 M * litage daniel_hozac: should i add the ip address while the guest is running or stop? 1170890712 M * xp_prg2 Not enough information: "dev" argument is required. 1170890734 M * daniel_hozac litage: doesn't matter. 1170890739 M * daniel_hozac xp_prg2: so, you need dev as well. 1170890758 M * xp_prg2 can you show me an example please? 1170890794 Q * ensc Ping timeout: 480 seconds 1170890808 M * daniel_hozac ip addr del
/ dev ethX? 1170890837 M * xp_prg2 yes I think that is right, trying 1170890998 M * xp_prg2 that worked, 1170891003 M * litage daniel_hozac: i added the problem-guest's ip address to the vserver host, then started the guest 1170891040 M * xp_prg2 what is wierd is it is not resolving dns names but I can ping out 1170891054 M * xp_prg2 I have another centos4 vserver and it is pinging out fine and resolving domain name 1170891056 M * xp_prg2 s 1170891066 M * xp_prg2 any ideas? 1170891132 M * daniel_hozac did you configure /etc/resolv.conf? 1170891137 M * daniel_hozac litage: and? 1170891153 M * xp_prg2 I don't have a resolv.conf in my other centos4 and it works ok 1170891157 M * Bertl xp_prg2: double check that ping -I www.google.com works on the host 1170891186 M * daniel_hozac if you don't have a resolv.conf file, nothing is going to be resolved. 1170891190 M * litage daniel_hozac: and the vserver host just lost the problem-ip ... ``ip addr'' doesn't list it any longer 1170891196 M * xp_prg2 yes that works 1170891200 M * xp_prg2 ok I will add a resolv.conf 1170891207 M * daniel_hozac litage: and if you add it again? 1170891294 M * litage daniel_hozac: things seem to be working...just ssh'd into the problem-guest 1170891315 M * daniel_hozac so for some reason, vserver ... start is removing the address? 1170891335 M * litage daniel_hozac: not sure 1170891335 M * xp_prg2 that fixed it thanks! 1170891392 M * litage daniel_hozac: ``ip addr'' on the host lists the problem-guest's ip address differently from the others though: http://rafb.net/p/ShH14U56.html 1170891417 M * daniel_hozac yes, you didn't specify a broadcast address. 1170891496 M * litage daniel_hozac: does that matter? 1170891503 M * litage (at the moment?) 1170891509 M * daniel_hozac no. 1170891545 J * rgl ~Rui@84.90.10.107 1170891556 M * litage k 1170891598 M * litage going to stop the problem-guest while its ip address is bound on the host to see if stopping gets rid of the address 1170891611 M * daniel_hozac it does. 1170891641 M * litage yup 1170891665 M * daniel_hozac could you paste a vserver --debug start run? 1170891706 M * litage i just started the problem-guest. part-way through starting up, ``ip addr'' on the host listed the problem-guest's ip address. but once the guest had finished starting, ``ip addr'' no longer listed the address 1170891719 M * litage daniel_hozac: ``vserver --debug start run" or without the "run"? 1170891729 M * daniel_hozac no run. 1170891737 J * nou Chaton@causse.larzac.fr.eu.org 1170891739 M * litage k 1170891762 M * xp_prg2 what is the best way to remove a server? 1170891764 Q * debugger Ping timeout: 480 seconds 1170891767 M * daniel_hozac vserver ... delete 1170892180 M * litage daniel_hozac: here's the output from using --debug: http://paste.debian.net/21574 1170892267 M * daniel_hozac it says it fails to start your guest. 1170892273 M * daniel_hozac so of course the address is removed. 1170892297 M * daniel_hozac at that point, it starts doing cleanup. 1170892387 M * daniel_hozac what distro is that? 1170892391 M * litage daniel_hozac: debian sarge 1170892509 M * daniel_hozac might want to try appending exit 0 to that. 1170892519 M * litage daniel_hozac: to what? 1170892539 M * daniel_hozac to the rc script, /etc/init.d/rc. 1170892559 M * litage daniel_hozac: in the host or guest? 1170892571 M * daniel_hozac guest. 1170892737 M * litage daniel_hozac: ``diff /vserver/golive/etc/init.d/rc /vserver/apt/etc/init.d/rc'' succeeds, and the 'apt' guest works 1170892757 M * daniel_hozac but i assume it runs different services? 1170892763 M * litage ah true 1170892781 M * litage daniel_hozac: is it possible to start a guest up in runlevel 1? 1170892789 M * daniel_hozac sure.