1173226068 Q * derjohn Ping timeout: 480 seconds 1173226071 J * derjohn ~derjohn@dslb-084-058-193-251.pools.arcor-ip.net 1173226452 Q * ZLinux Remote host closed the connection 1173226965 Q * rob-84x^ Ping timeout: 480 seconds 1173230928 Q * meandtheshell Quit: Leaving. 1173232112 M * jkl it seems as if a guest has access to all networks a host is connected to, regardless of whether it has an interface assigned to it on a particular network 1173232127 M * jkl is there a way to prevent this? 1173232284 M * jkl guest#> ip route get 192.16.1.10 1173232293 M * jkl 192.168.1.10 dev if3 src 192.168.1.1 cache mtu 1500 advmss 1460 hoplimit 64 1173232298 M * jkl i have never seen anything like that before 1173232320 M * jkl (the host has 192.168.1.1) 1173232773 Q * softi42 Ping timeout: 480 seconds 1173233385 J * softi42 ~softi@p549D69A2.dip.t-dialin.net 1173237120 J * dmax ~semaj@81.193.56.8 1173237983 J * _jthm_ ~infowolfe@c-67-164-195-129.hsd1.ut.comcast.net 1173238132 J * daniel_hozac|irssi ~daniel@c-091472d5.010-230-73746f22.cust.bredbandsbolaget.se 1173238231 Q * daniel_hozac Ping timeout: 480 seconds 1173238237 Q * infowolfe_ Read error: Connection reset by peer 1173240338 Q * ensc Killed (NickServ (GHOST command used by ensc_)) 1173240348 J * ensc ~irc-ensc@p54B4F1D3.dip.t-dialin.net 1173242555 J * Daniel15 ~daniel@c58-107-237-225.sunsh7.vic.optusnet.com.au 1173242588 M * Daniel15 Is it possible to change the name of a vserver? I have a vserver called "public01", and want to rename it to "gforge01" 1173242670 M * Daniel15 Don't worry, I found something on the old wiki (http://oldwiki.linux-vserver.org/TweakingTheConfig) :) 1173242766 N * daniel_hozac|irssi daniel_hozac 1173242797 M * Daniel15 One question... Are the symlinks in /etc/vservers/vserver_name hardlinks, or symlinks? 1173242810 M * Daniel15 I meant, are the *links* hardlinks or symlinks :P 1173242829 M * daniel_hozac symlinks, of course. 1173242844 M * daniel_hozac you can't have hardlinks to files on other filesystems or directories. 1173242872 M * Daniel15 Oh, OK (sorry, I'm still a noob at this stuff :P) 1173242879 M * Daniel15 Thanks :) 1173243058 J * hall ~hall919@218.206.101.202 1173243080 M * hall Why nobody speak? 1173243124 M * daniel_hozac because you've only been here for a minute? 1173243176 M * hall ^-^ 1173243191 M * hall only one minute! 1173243215 M * Daniel15 Another question, how do I properly change the hostname? 1173243231 M * Daniel15 I edited /etc/hostname, restarted the vserver, and the hostname is still set to the old one 1173243274 M * hall I'm a now user,don't know 1173243282 M * hall I'm a new user,don't know 1173243289 M * daniel_hozac is the initscript that sets it based on that file running? 1173243302 M * daniel_hozac usually you'd set it from /etc/vservers//uts/nodename. 1173243405 M * Daniel15 Thanks :). I edited that file, and it works fine 1173243674 Q * fs charon.oftc.net neutron.oftc.net 1173243674 Q * click charon.oftc.net neutron.oftc.net 1173243674 Q * Hunger charon.oftc.net neutron.oftc.net 1173243687 J * fs fs@213.178.77.98 1173243709 M * hall ×ßÁË£¬Ð»Ð» 1173243713 P * hall 1173243748 J * click click@ti511110a080-0949.bb.online.no 1173243804 J * Hunger Hunger.hu@Hunger.hu 1173243845 Q * weasel Ping timeout: 480 seconds 1173243863 Q * pisco_ Ping timeout: 480 seconds 1173243897 Q * Piet Ping timeout: 480 seconds 1173244582 J * DoberMann_ ~james@AToulouse-156-1-167-147.w90-38.abo.wanadoo.fr 1173244690 Q * DoberMann[ZZZzzz] Ping timeout: 480 seconds 1173245095 Q * Daniel15 Ping timeout: 480 seconds 1173247728 J * gab ~gab@158.36.45.236 1173248305 Q * gwar Ping timeout: 480 seconds 1173249379 J * SNy_ 0d3ca6b4c6@bmx-chemnitz.de 1173249379 Q * SNy Read error: Connection reset by peer 1173249394 N * SNy_ SNy 1173251663 Q * sladen Ping timeout: 480 seconds 1173251913 Q * derjohn Ping timeout: 480 seconds 1173252667 N * DoberMann_ DoberMann 1173253346 Q * ensc Killed (NickServ (GHOST command used by ensc_)) 1173253356 J * ensc ~irc-ensc@p54B4F1D3.dip.t-dialin.net 1173253375 Q * softi42 Ping timeout: 480 seconds 1173253953 J * softi42 ~softi@p549D69A2.dip.t-dialin.net 1173255435 J * sladen paul@starsky.19inch.net 1173255688 N * DoberMann DoberMann[PullA] 1173255881 J * dna ~naucki@p54BCF1F6.dip.t-dialin.net 1173256379 N * Bertl_oO Bertl 1173256385 M * Bertl morning folks! 1173256404 M * Bertl jkl: hmm? please elaborate ... 1173256443 M * Bertl sannes: ping! 1173256555 J * dlezcano_ ~dlezcano@blueice4n2.uk.ibm.com 1173257916 N * DoberMann[PullA] DoberMann 1173257952 J * weasel weasel@asteria.debian.or.at 1173258032 M * Bertl wb weasel! hey DoberMann! 1173258123 M * DoberMann morning Bertl 1173258833 J * dna_ ~naucki@p54BCCCD1.dip.t-dialin.net 1173259138 Q * dna Ping timeout: 480 seconds 1173260109 J * meandtheshell ~markus@85-124-206-87.dynamic.xdsl-line.inode.at 1173260177 Q * shedi Quit: Leaving 1173260203 J * ema ~ema@lart.galliera.it 1173260319 Q * Adrinael Quit: s|g 1173260348 J * Adrinael adrinael@st12-127.tky.hut.fi 1173260733 M * matti Bertl: :) 1173260748 M * Bertl hey matti! 1173260901 M * matti How are you? 1173260930 M * Bertl fine, tx ... and you? 1173260936 J * lilalinux ~plasma@80.69.41.3 1173261132 M * matti Not bad. 1173261152 M * matti :) 1173262757 Q * cdrx Quit: Leaving 1173262765 Q * SNy Ping timeout: 480 seconds 1173263022 J * SNy b8de331dfb@bmx-chemnitz.de 1173263040 J * Machuidel ~mvierssen@e22200.upc-e.chello.nl 1173263078 M * Bertl welcome Machuidel! 1173263084 M * Machuidel Hi 1173263158 M * Machuidel Anyone knows how to share a single IP with multiple vservers. Of cource I will not start services on the same port twice. 1173263161 M * Machuidel ? 1173263201 M * Machuidel I knew how to do it in vserver1, but things changed :) 1173263337 M * Bertl did they? 1173263360 M * Bertl well, you basically assign the ip to several guests ... that's 5t 1173263386 M * Bertl *it 1173263412 J * derjohn ~derjohn@80.69.41.3 1173263458 M * Machuidel yes. they did :). You could then set IPROOT to the same IP as the root server 1173263480 M * Machuidel I'll will try assigning the same address to multiple vservers 1173263496 M * Bertl except for the new style config (which is now 3 years old :) nothing changed in that regard 1173263529 M * Bertl you want to use the nodev feature though, as you do not want the tools to create the ip for you 1173263611 M * Machuidel okay. thank you. Now I know what went wrong when I lost the connection to the root server. The vserver took over the primary IP :P 1173263651 M * Bertl well, yes, probably more problematic was that the tools will have taken down the ip on guest shutdown 1173263692 M * Machuidel yep. That was indeed the moment I lost the connection to the root server :P 1173263741 M * Bertl but you should note that sharing the host ip with a guest is somewhat dangerous 1173263754 M * Machuidel So since then I just started masquerading to internal vserver IP's 1173263760 M * Bertl and in many cases can be avoided completely 1173263781 M * Machuidel Yeah. If someone is going to install SSH it will be problematic 1173263803 M * Bertl yes, especially as he can probably DoS your host system 1173263808 M * Machuidel What is the best way to avoid it? 1173263833 M * Bertl use private IPs for the guests, and S/DNAT the 'known good' ports to them 1173263862 M * Bertl this allows you precise control over the ports on the host 1173263873 M * Bertl while it doesn't harm operation for the guests 1173263997 M * Machuidel Sounds as the best solution. I have been doing it that way for a while and works pretty well. Will just keep it like that. 1173264024 M * Machuidel The risk is greater than the comfort :) 1173264046 M * Bertl yes, the thing is a little different if you have two public ips 1173264057 M * Bertl and share one of them between all guests 1173264085 M * Bertl might not be a friendly scenario for the different guests, but is secure from the host PoV 1173264110 M * Machuidel That is like my situation. I've only got 2 public IP's 1173264204 M * Machuidel At least as the host is safe it will be good enough for me ;) 1173264213 M * Bertl so you could go that road, if you do not care what the guests do to eachother, and do not want to bother with the S/DNAT 1173264284 M * Machuidel Yep. S/DNAT will do for me. Thank you for helping. 1173264293 M * Bertl you're welcome! 1173264305 M * Machuidel cya ;) 1173264314 M * Bertl feel free to hang around ... 1173264343 M * Machuidel tnx 1173264384 P * Machuidel Vserver IP sharing enlightened 1173265965 M * bXi HELP 1173265972 M * bXi vservers wont start anymore :( 1173265996 M * Bertl ubuntu? 1173266015 M * bXi gentoo 1173266022 M * bXi i updated everything 1173266032 M * Bertl unusual ... what do they complain about on startup? 1173266034 M * bXi and now its whining about WARNING: can not access /proc/uptime 1173266045 M * bXi but i thought the /etc/init.d/vserver did this 1173266052 M * bXi but they changed that into vserver.default as well 1173266055 M * Bertl you didn't run the vprocunhide then 1173266072 M * Bertl (which should be run on system startup once) 1173266176 M * bXi ah it works 1173266188 M * bXi but isnt this the old way again? 1173266209 M * Bertl didn't change for the last two years 1173266240 M * bXi i now have 2 init files instead of 1 again 1173266249 M * bXi maybe i had some old obscure version 1173266258 M * bXi but the vservers are up again 1173266481 M * bXi thanks 1173266492 M * Bertl you're welcome! 1173267158 M * meandtheshell hihi #vserver aka #emergency_room :) 1173267180 M * Bertl :) 1173270062 M * renihs narf :) 1173270638 M * Bertl nap attack ... back later ... 1173270643 N * Bertl Bertl_zZ 1173271070 Q * Hollow Ping timeout: 480 seconds 1173271406 J * marl ~matt@84.92.193.226 1173271538 M * marl hi can anyone point me in the rite direction asto what i've mucked up on this? i have jsut installed ubuntu edgy, and the vserver kernels from u32, 2.6.17-11-vserver-server, and have tried creating a vserver guest, but i get hte following error at the end of the build script : W: Failure trying to run: chroot /etc/vservers/.defaults/vdirbase/mysql1 mount -t proc proc /proc 1173271583 J * Hollow ~hollow@styx.xnull.de 1173271585 P * Hollow 1173272507 J * ZLinux ~ZLinux@88.213.57.11 1173273305 J * zagor ~andrea@service.cab.unipd.it 1173274480 J * marcfiu ~mef@aegis.CS.Princeton.EDU 1173274654 P * zagor I'm gonna save Rose Tyler from the middle of the Dalek fleet, and then I'm gonna save the Earth 1173275744 J * Hollow ~hollow@styx.xnull.de 1173279232 Q * mjt Read error: Connection reset by peer 1173279275 J * mjt ~mjt@nat.corpit.ru 1173279454 Q * mjt Read error: Connection reset by peer 1173279499 Q * gab Quit: Leaving 1173279558 J * mjt ~mjt@nat.corpit.ru 1173279570 J * mjt_ ~mjt@nat.corpit.ru 1173279585 P * mjt_ 1173282190 Q * Johnnie Remote host closed the connection 1173282333 J * Johnnie ~jdlewis@jdlewis.org 1173282700 P * marcfiu 1173283208 Q * renihs Ping timeout: 480 seconds 1173283220 J * cdrx ~legoater@blueice1n1.uk.ibm.com 1173283345 J * fallingdutch ~fallingdu@krlh-590f25d8.pool.einsundeins.de 1173284322 N * Bertl_zZ Bertl 1173284390 M * Bertl marl: could you paste the full message to paste.linux-vserver.org (it seems cut off) 1173284446 M * Bertl welcome fallingdutch! 1173284491 M * fallingdutch thank you Bertl 1173284587 M * fallingdutch Bertl, on the wiki there is mentioned a secure logging system for all vserver to the "main" system, but never described in detail - do you know how to do that? 1173284631 J * renihs ~penguin@83-65-34-34.arsenal.xdsl-line.inode.at 1173284634 M * fallingdutch found a solution when searching the web but maybee that solution is better 1173284692 M * Bertl fallingdutch: you mean system logs, yes? 1173284716 M * fallingdutch yes 1173284752 M * Bertl I would go for syslog-ng then, and either put the sockets into the guests, or use 'network' logging form a guest ip to the host's 'logging ip' 1173284776 M * Bertl both is probably roughly the same speed, and syslog-ng can handle it quite nicely 1173284795 M * fallingdutch found a bit diffrent by using ssh and syslog 1173284818 M * Bertl that for sure adds some overhead, but probably works fine too 1173284855 M * fallingdutch so there seems not to be a special vserver solution - just normal network logging, right? 1173284891 M * Bertl yes, system logging is userspace, and userspace is supposed to work quite fine, that's why there is no 'special' solution :) 1173284901 M * fallingdutch ;) 1173284903 M * fallingdutch ok, thanks 1173284906 M * Bertl np 1173284923 M * fallingdutch do you happen to know how to enable a bridge in /etc/network/interfaces? 1173284959 M * Bertl there are bridgeutils whic can be used to do that for you 1173284974 M * fallingdutch i know, but they will be down after reeboot - afaik 1173284994 M * Bertl ah, i.c. what you mean ... 1173285015 M * Bertl there was a config file for that, but this is probably very distro specific 1173285036 M * mjt wug. Only pppd & vtund are left on the host.... :( 1173285041 M * fallingdutch ok ... sounds like scripting ;) 1173285083 M * Bertl fallingdutch: I'd suggest to google for that and see what's already there 1173285101 M * fallingdutch Bertl, haven't found any usefull so far 1173285108 M * fallingdutch thats why i am asking here 1173285117 M * Bertl what distro? 1173285120 M * fallingdutch debian 1173285123 M * Bertl mjt: is that bad? 1173285133 M * Bertl debian ethernet bridge setup 1173285144 M * Bertl (would be my search words) 1173285149 M * fallingdutch mine, too 1173285165 M * fallingdutch setting them up is not a problem and there are hundreds of howtos etc 1173285183 M * fallingdutch the problem is setting them up at boot-time 1173285211 M * mjt sure it's... bad ;) Because the two should work together, and I don't know how to grant serial port access to vserver. 1173285245 M * Bertl fallingdutch: wait for micah or waldi, they might know more ... 1173285270 M * fallingdutch thank you, Bertl 1173285271 M * Bertl mjt: that is quite easy, you copy the serial device into the guest, that's it 1173285333 M * Bertl mjt: but for the pppd, you will require raised capabilities I guess, because it for sure wants to setup the ppp interfaces 1173285345 M * mjt it is 1173285372 M * mjt which is.. not good... 1173285403 M * Bertl well, it is still saver than on the Host or on any other Linux system :) 1173286156 M * waldi fallingdutch: what do you want to know? 1173286180 M * Bertl automatic boot time bridge setup in debian :) 1173286227 M * waldi read /usr/share/doc/bridge-utils/README.Debian.gz 1173286678 J * _dmax ~semaj@81.193.63.119 1173287035 Q * dmax Ping timeout: 480 seconds 1173287036 N * _dmax dmax 1173287710 Q * cdrx Read error: Connection reset by peer 1173288635 Q * Roey Ping timeout: 480 seconds 1173289827 M * derjohn Hi, I converted the schedule file to the new sched dir format. But it seems the guest is "stuck" "hogged" to the 1st CPU. do I have to enable the 2nd CPU support? 1173289848 M * daniel_hozac hmm? 1173289861 M * Bertl derjohn: hmm, interesting ... hello btw! 1173289871 M * derjohn in "top" one cpu is at 100%, the second at <= 1% .... 1173289884 M * daniel_hozac so you have two cpuhogs, yes? 1173289888 M * Bertl derjohn: more than one process inside the guest? 1173289889 M * derjohn I use a (dont lart me ) 2.6.18.5 2.1.1.3 1173289907 M * derjohn Bertl, yes, a full webserver env ... Bertl hello BTW :) 1173289925 M * daniel_hozac what does find /etc/vservers//sched | xargs tail output? 1173289938 M * daniel_hozac and contents of /proc/virtual//sched 1173289984 M * derjohn boerteltrand:~# cat /etc/vservers/gerstung3/sched/* |xargs 1173289984 M * derjohn 30 32 1000 200 1173290000 M * daniel_hozac hmm? that's not what i said. 1173290008 M * derjohn i didnt want to flood ;) 1173290016 M * daniel_hozac paste.linux-vserver.org, please. 1173290018 M * Bertl btw, completely unrelated, the check/repair added in 2.6.17/18 for softraid is quite nice :) 1173290073 M * derjohn daniel_hozac, http://paste.linux-vserver.org/1242 1173290121 M * derjohn argh ! 1173290123 J * shedi ~siggi@ftth-237-144.hive.is 1173290125 M * derjohn root 16686 99.1 0.0 3056 1016 ? R< Mar06 1972:46 \_ /sbin/modprobe -q -- net_pf_10 1173290152 M * derjohn there is some modprobe running at 100% .. that is obviously only on only one CPU 1173290217 M * derjohn why does [khelper] start a modprobe ? A Kernel Thread that runs a userspace prog ? 1173290234 M * Bertl udev/hotplug 1173290243 M * daniel_hozac that's basically khelper's job. 1173290264 M * daniel_hozac (starting userspace programs on behalf of the kernel) 1173290278 M * derjohn hm, kann I kill the modprobe without danger ? 1173290306 M * daniel_hozac probably 1173290312 M * derjohn -9 1173290314 M * derjohn ? 1173290317 M * daniel_hozac right. 1173290333 M * derjohn it's kill resistant :) 1173290350 M * derjohn time to reboot 1173290394 M * derjohn Bertl, and mdadm on Debian also supports a monthly md-check via cron --- 1173290402 M * derjohn i mean the mdadm package 1173290416 M * Bertl I would check _why_ the modprobe did hang 1173290436 M * Bertl because it is likely to get triggered on system startup 1173290441 M * derjohn how so ? 1173290446 M * derjohn syslog ? 1173290472 M * derjohn its parameters were "net_pf_10" 1173290475 M * Bertl e.g., and/or modprobe config 1173290499 M * derjohn well, I leaving tomorrow at 10 for holiday till sunday ... 1173290512 M * derjohn so i will do further investigation afterwards 1173290519 M * derjohn (after next hang) 1173290549 N * DoberMann DoberMann[PullA] 1173290555 M * Bertl what holiday do you celebrate? 1173290593 M * derjohn nothing special ... just with "upsteam family" 4 days of skiing in .CH 1173290608 M * derjohn (I just found a fon hotspot near the hotel :)) 1173290754 M * daniel_hozac derjohn: btw, you know you've given that guest ~93% of the CPUs, right? 1173290811 M * derjohn daniel_hozac, the idea was to be able to login on the host via ssh if amavis/spamassassin or whatever eats up all resources of the guest 1173290826 M * derjohn there is only one guest on that host 1173290830 M * daniel_hozac ok, just checking. 1173290836 M * derjohn daniel_hozac, thx yes. 1173290845 M * derjohn maybe I rise to 31/32 ..... 1173290875 M * derjohn the only thing on the host is sshd and mdadm (for raid fail notify) 1173290902 M * derjohn but: I saw there was something new woth cpusets and idle-time advancing 1173290911 M * derjohn i mean, in the scheduler 1173290948 M * daniel_hozac the scheduler config? 1173290961 M * daniel_hozac because the idle time has been around since 2.1.0 or so, IIRC. 1173291129 M * derjohn yup, in former times there were only 5 lines in the file. 1173291153 M * derjohn so, there was no config option for idle_time in the scheduler 1173291233 M * Guy- uh, I just noticed something strange that doesn't have to do anything with vserver, but I don't know who else to ask :) I have a mysqld process with two threads. The process is a zombie, but one of the threads is consuming 99% CPU. How can this be? Please do tell me to take my off-topic ramblings elsewhere... 1173291546 M * daniel_hozac are both threads dead? 1173291555 M * Guy- 'dead' as in what? 1173291561 M * daniel_hozac zombies. 1173291571 M * Guy- no, the one with 99% CPU usage isn't 1173291586 M * Guy- at least that's what I assume, because a zombie can't use the CPU 1173291597 M * Guy- but I can't kill it, kill -9 doesn't do anything 1173291613 M * daniel_hozac anything in dmesg? 1173291618 M * Guy- no 1173291634 M * Guy- am I right in assuming this isn't supposed to happen? 1173291665 M * Guy- I'm still a bit unclear on multithreading, but I didn't think a multithreaded process could become a zombie (i.e. exit) while it still had active threads 1173291685 M * Guy- I mean, when the process exited, all threads should have exited too, shouldn't they? 1173291691 M * Bertl it should still be possible to send signals to it 1173291705 M * Guy- well, kill succeeds 1173291717 M * Bertl so that is fine then, no? 1173291718 M * Guy- but even the KILL signal has no effect 1173291729 M * daniel_hozac does strace show anything? 1173291735 M * Guy- no, it just sits there 1173291735 M * Bertl Guy-: ah, you mean, it doesn't succeed :) 1173291749 M * Guy- Bertl: it succeeds (doesn't return an error), but doesn't do anything useful either :) 1173291761 M * Bertl try specifying - 1173291795 M * Bertl (might be a little trickier with normal kill, e.g. use -- ) 1173291808 M * Guy- you mean, send a signal to the process group? 1173291813 M * Bertl yep 1173291879 M * Guy- OK, thanks (could've thought of that myself...); I'll have the person whose box this is try it 1173291892 M * Bertl make that, and let us know ... 1173291896 M * Guy- what's the deal with strace and multithreaded processes, btw? 1173291906 M * daniel_hozac ? 1173291912 M * Guy- when you strace a multithreaded process, strace won't exit on ctrl-c 1173291912 M * Bertl works fine, if you specify the proper options 1173291915 M * Guy- ah 1173291922 M * daniel_hozac ah, -fF? 1173291934 M * Guy- I always use -ffF 1173291944 M * Bertl that should be fine then ... 1173291948 M * Guy- it isn't 1173291970 M * Guy- it goes 'detaching PID', and then just sits there until I kill it 1173291984 M * Guy- sometimes it leaves the traced process in the T state so I have to send it a CONT 1173291990 M * Bertl well, maybe your kernel is not really working as expected 1173291998 M * daniel_hozac yeah, that sounds rather broken. 1173292001 M * daniel_hozac what kernel is that? 1173292012 M * Guy- 2.6.19.2-vs2.2.0-rc8.7 1173292025 M * Guy- but I've seen this behaviour with any kernel, really 1173292038 M * Guy- it's fine if I start the process straced 1173292052 M * daniel_hozac you sure it's not stuck in userspace? 1173292063 M * Guy- but if I attach to a running process that has several threads, it won't exit 1173292064 M * matti Bertl, daniel_hozac :) 1173292068 M * daniel_hozac hey matti. 1173292070 M * Guy- daniel_hozac: how do you mean? 1173292120 M * daniel_hozac e.g. in a for (;;) ; loop. 1173292129 M * daniel_hozac or something that doesn't call into the kernel 1173292151 M * Guy- the mysql process? I'm almost certain it's stuck in userspace, that's why I don't see a thing in strace 1173292163 M * Guy- but I don't see how strace could become stuck when I press ctrl-c 1173292178 M * daniel_hozac how are you stracing it? is the mysqld in a guest? 1173292187 M * Guy- no, it's not even a vserver kernel 1173292208 M * Guy- I just asked here because this channel has the highest concentration of kernel knowledge of all channels I'm on 1173292218 M * daniel_hozac so the vs2.2.0-rc8.7 above was just for show? :) 1173292228 M * Guy- no, these are two different computers :) 1173292248 M * Guy- I get the 'strace hangs when I press ctrl-c while stracing a multithreaded process' on the one with vs2.2.0-rc8.7 1173292277 M * Guy- and the 'one unkillable thread of zombie mysqld eats CPU' issue is on a different box 1173292278 M * daniel_hozac only for multithreaded processes? 1173292282 M * Guy- yes 1173292289 M * daniel_hozac that's at least strange. 1173292298 M * Guy- but I can reproduce it with a vanilla kernel too 1173292299 M * daniel_hozac could be the reparenting bug though, i suppose. 1173292308 M * daniel_hozac ok, so that's not it :) 1173292308 M * Guy- so I don't think it has to do with vserver specifically 1173292342 M * Guy- to me this has just been the way strace worked, really :) I knew it was a bad idea to attach it to multithreaded processes, but didn't really think about it 1173292349 M * Guy- and now you're telling me it should work? :) 1173292357 M * daniel_hozac i don't see why not. 1173292363 M * Guy- can you try? 1173292396 M * Guy- attach strace to, say, a mysqld thread 1173292402 M * Guy- and then hit ctrl-c after a few seconds 1173292434 M * Guy- strace hangs in do_wait 1173292456 M * daniel_hozac so, stupid question time, how do you attach to a thread? i mean, how do you get the tid, and how do you specify that? 1173292466 M * Guy- ps -eLf 1173292473 M * Guy- the LWP column has the thread id 1173292479 M * Guy- and you specify it like you would a process ID 1173292509 M * Guy- i.e. strace -ffF -p TID 1173292656 M * Guy- if I strace the hanging strace, I see its wait4()ing for a different thread of the same process (not the one I attached it to) 1173292729 M * daniel_hozac hmm, ok, that sounds like a bug in strace then. 1173292741 M * Guy- OK, that's what I thought 1173292762 M * Guy- but you were giving me the impression this was something you used without problems all the time :) 1173293096 M * daniel_hozac not at all. 1173293713 Q * fallingdutch Ping timeout: 480 seconds 1173293973 J * marcfiu ~mef@aegis.CS.Princeton.EDU 1173293982 M * Bertl welcome marcfiu! 1173293985 M * marcfiu hey Bertl 1173294004 M * marcfiu this just came up: what's involved with running X out of a vserver vs. root context? 1173294019 M * Bertl hehe, depends on the X 1173294081 M * Bertl for example, if you use vnc as X server, then this doesn't need anything 1173294094 M * Guy- how about Xorg or Xfree86? 1173294097 M * Bertl if you use a framebuffer based X, you will need access to the gb devices 1173294104 M * Bertl *fb 1173294131 M * Bertl if you want to run X with hardware access, you might need to give iopl access too 1173294181 M * marcfiu want to run something like mythtv out of a vserver. 1173294196 M * marcfiu So my guess is that it would require hardware access. 1173294226 J * stefani ~stefani@flute.radonc.washington.edu 1173294234 M * Bertl mythtv needs access to the dvb/tv devices 1173294247 M * Bertl which is simple a question of copying them into the guest 1173294277 J * Piet hiddenserv@tor.noreply.org 1173294293 M * Bertl welcome stefani! Piet! 1173294302 M * stefani aloha 1173294368 M * phedny marcfiu: I just ordered a dvb-c pci card that I plan to use from within a vserver 1173294389 M * phedny marcfiu: it'll be delivered Monday, if you like I can report my success / failure to you? 1173294396 M * Guy- Bertl: any idea what capabilities Xorg would need? 1173294413 M * daniel_hozac http://oldwiki.linux-vserver.org/MoreUbuntu 1173294415 M * Guy- (obviously I can try and err) 1173294418 M * Piet grüzi alle miteinadr 1173294442 M * Bertl Guy-: I would start without giving special caps first 1173294483 M * Guy- daniel_hozac: ah, you already showed me this once, I remember now 1173294491 M * Bertl then add SYS_RAWIO and maybe SYS_MODULE 1173294509 N * DoberMann[PullA] DoberMann 1173295629 M * marcfiu phedny: please do 1173295945 M * cehteh Bertl: apropos moving guests http://cryopid.berlios.de/ just a note if you dont know it, cant handle that but might be partially useful 1173295947 J * rob-84x^ ~rob@submarine.ath.cx 1173296004 M * Bertl cehteh: interesting indeed ... 1173296040 M * cehteh Bertl: actually i would like if most of this stuff can be done in userland 1173296111 M * cehteh problems are still stateful servers like X and states which are only maintained in the kernel .. but for snapshoting simple servers its already enough 1173296397 J * user1 ~user1@189.131.131.155 1173296401 N * user1 Term_ 1173296411 M * Bertl welcome Term_! 1173296421 M * Term_ hi bert 1173296481 M * meandtheshell Bertl: do you use emacs? 1173296489 M * Bertl meandtheshell: nope 1173296493 M * daniel_hozac vim! 1173296493 M * meandtheshell I see 1173296502 A * Bertl is one of the vim folks :) 1173296504 M * cehteh would be the first kernel hacker who does :) 1173296522 M * meandtheshell cehteh: what - using emacs? 1173296526 A * cehteh uses emacs .. but i rarely do kernel hacking :) 1173296530 M * cehteh meandtheshell: yes 1173296541 M * cehteh well execpt linux who uses uemacs 1173296546 M * cehteh linus 1173296559 M * meandtheshell cehteh: hm ... really 1173296580 M * meandtheshell daniel_hozac: you are with vim too? 1173296586 M * daniel_hozac yes. 1173296594 M * meandtheshell ok 1173296610 M * cehteh mhm for 1. April i wanted to prepare a EmacsOS live CD ... linux kernel with init=/bin/emacs ... 1173296615 M * cehteh maybe 2008 :) 1173296618 M * daniel_hozac haha. 1173296625 M * daniel_hozac great idea. 1173296663 M * cehteh only emacs and nothing else :) 1173296673 M * cehteh maybe X ... 1173296693 M * meandtheshell vmemacs loading linux.elc :) 1173296694 M * cehteh there is even a X windowmanager written in elisp 1173296721 M * cehteh emacs.ko ;) 1173296957 J * dreamind ~dreamind@C2107.campino.wh.tu-darmstadt.de 1173296979 M * Bertl wb dreamind! 1173296985 M * dreamind Hi Bertl :) 1173296989 M * dreamind Hi folks 1173297128 J * jmcaricand ~kvirc@d90-144-26-67.cust.tele2.fr 1173297140 M * Bertl welcome jmcaricand! 1173297151 M * jmcaricand Hello bert 1173297177 M * dreamind Bertl: hm, do you have some kind of auto-greet-script ;) 1173297192 M * Bertl dreamind: nope 1173297201 M * dreamind ok, then :D 1173297217 M * dreamind had such a thing some time ago, when I used xchat on linux ;) 1173297228 M * dreamind but only limited to some channels and some people ;) 1173297348 M * jmcaricand I want to install amavis (with fsecure av) in a guest but amavis use 127.0.0.1 interface. Does somebody use amavis in a guest ? 1173297375 M * daniel_hozac what's the problem with it? 1173297380 M * Bertl jmcaricand: well, if it uses 127.0.0.1 it is broken 1173297393 M * Bertl jmcaricand: it should use 'localhost' instead 1173297412 M * daniel_hozac but even if it is broken, it should work (unless it's horribly broken). 1173297420 M * Bertl yep 1173297482 M * jmcaricand I will test again tomorrow. 1173297678 M * Bertl okay, make that and please let us know how it goes 1173297910 M * jmcaricand Bertl: OK 1173297912 Q * phreak`` Quit: leaving 1173297988 J * phreak`` ~phreak``@deimos.barfoo.org 1173298066 M * PowerKe I run amavis on a guest at work 1173298199 M * jmcaricand PowerKe: You use 127.0.0.1 interface or other ? 1173298274 M * PowerKe Usually I use the private IP 1173298300 M * PowerKe (my vpn seems to be down since my last kernel upgrade) 1173298599 P * marcfiu 1173298654 M * PowerKe Ok, got it running again 1173298772 M * Hollow daniel_hozac: ping 1173298777 M * daniel_hozac pong 1173298819 Q * jmcaricand Quit: KVIrc 3.2.4 Anomalies http://www.kvirc.net/ 1173298833 M * PowerKe Seems like I have localhost and the private IP in the config 1173298845 M * Hollow i tried the new scheduler config today, works great, except if i only touch idle-time it complains about not a number .. though the pure existance of the file should already be enough, no? 1173298861 M * PowerKe ping still resolves localhost to 127.0.0.1 though... 1173298864 M * daniel_hozac humm, yes... 1173298889 M * daniel_hozac so touching idle-time doesn't work? 1173298972 M * daniel_hozac hmm, true. how on earth did i miss that? 1173299661 M * daniel_hozac ok, should be fixed now. 1173299867 M * daniel_hozac thanks. 1173300437 Q * tamitall Read error: Connection reset by peer 1173300776 Q * Term_ Quit: ircII2.8.2-EPIC3.004 --- Bloatware at its finest. 1173301160 Q * ema Quit: leaving 1173301859 M * Hollow daniel_hozac: thanks :) 1173302070 Q * meandtheshell Quit: Leaving. 1173302255 M * Guy- Bertl, daniel_hozac: gdb refused to attach to the runaway mysqld thread, saying 'no such file or directory', and killing the process group had no effect; alas, the box had to be rebooted so no further experiments are possible 1173302277 M * Bertl very interesting ... 1173302331 M * Guy- Bertl: but if it happens again, I'll try to get the admin to play with it some more and tell you what we find, if you're interested 1173302356 M * Bertl sure, sounds like a kernel (probably mainline) issue to me 1173302377 M * Guy- that's my guess too; this is 2.6.16.29 vanilla 1173302390 M * Guy- they rebooted into 2.6.20 now 1173302392 M * Bertl but just to make sure, you should upgrade to the latest versions 1173302396 M * Bertl ah, good 1173303195 J * bonbons ~bonbons@83.222.38.146 1173303211 M * Bertl hey bonbons! 1173303256 M * bonbons hey Bertl! 1173303481 J * marl_ ~matt@82-40-218-233.cable.ubr01.dunb.blueyonder.co.uk 1173303624 M * Bertl wb marl_! 1173303910 Q * marl Ping timeout: 480 seconds 1173304218 J * FireEgl Proteus@2001:5c0:84dc:1:211:9ff:feca:b042 1173305052 Q * bonbons Quit: Leaving 1173305404 P * stefani I'm Parting (the water) 1173306421 N * DoberMann DoberMann[ZZZzzz] 1173306862 J * er ~yakker@aegis.CS.Princeton.EDU 1173306878 M * Bertl welcome er! 1173306890 M * er Bertl, Hi Herbert 1173306933 M * er I've been playing with the latest vserver kernel + Eric's L2 virt patches 1173306946 M * Bertl ah, great! 1173306979 M * er Bertl, have you been trynig to merge his patches in as well? 1173306990 M * Bertl nope, not yet 1173307303 M * er does setting CAP_NET_ADMIN and other capabilities for guests depend on init having CAP_SETPCAP? 1173307353 M * Bertl no 1173307353 M * Bertl you basically add that to the bcapabilities 1173307512 M * Bertl all processess will then be allowed to obtain it 1173307512 M * Bertl either via suid or via capability transfer 1173307512 M * er oh ok, so guest processes don't have it by default 1173307512 M * Bertl nope 1173307512 M * Bertl http://linux-vserver.org/Capabilities_and_Flags 1173307925 Q * dreamind Quit: dreamind 1173308064 Q * micah Read error: Connection reset by peer 1173308071 J * micah ~micah@micah.riseup.net 1173308465 M * er Bertl, hm. doesn't seem to document how processes obtain capabilities, once the guest has been configured with them 1173308570 M * er so if I execute an suid'd executable, will the spawned process get the capabilities in /etc/vserver//bcapabilities? 1173309088 M * er ''The set of available system capabilities is inherited from the Linux kernel and applied to all processes contained within a context.'' 1173309140 M * er hm. that's probably the 'p' set. 1173310368 Q * dna_ Quit: Verlassend