1118796151 J * martin ~martin@host-138-38-227-118.nat.bath.ac.uk
1118796232 M * martin Have managed to get bot the traditional NFS user server and unfs3 running from a vserver (the version hsipped with Debian) - is this news?  I.E. is it worth posting to the wiki?
1118796435 M * Loki|muh everything is worthed to be documented
1118796439 M * Loki|muh -ed
1118796545 M * martin OK - any suggestions as to where?
1118796641 M * martin Hmmm... the NFS entry on the FAQ methinks
1118796693 M * Loki|muh http://linux-vserver.org/Documentation <-- and here
1118796900 N * Bertl_oO Bertl
1118796915 M * martin Morning Bertl
1118796920 M * Bertl evening folks! ;)
1118796954 M * Bertl hey martin! got around installing the kernel on an alpha system :)
1118796971 M * Bertl (means, I got around :)
1118796986 M * martin I'm still waiting for it to turn up but the owner has promised me that it will.
1118797008 M * Bertl kernel works fine, tools have some issues ...
1118797040 M * martin OK, have added note to the FAQ about user space NFS servers, if I have time to experiment more with NFS I'll write a full page about it.
1118797057 M * Bertl cool!
1118797060 M * martin Bertl: Cool, that makes... 5 architectures?
1118797072 M * Bertl hmm, let me check ...
1118797107 M * Bertl x86, x86_64, s390 and s390x (yes!), sparc64, sparc, ppc, alpha
1118797122 M * martin I stand corrected.
1118797131 M * martin Better get MIPS checked for you as well
1118797132 M * Bertl arm I did manage to boot the kernel, but userspace is missing
1118797151 M * Bertl a mipsel is waiting in my basement ...
1118797162 M * Bertl (got no time for that yet)
1118797171 M * martin Having had an ask around I /should/ be able to get m68k, mips and hppa, possibly arm
1118797182 M * martin s390x - nice
1118797188 M * Bertl ah, hppa I forgot, is working fine (basement too ;)
1118797215 M * martin OK, so it's now most architectures
1118797221 M * Bertl yeah, well, s390/x was on hercules, so no real hardware yet
1118797223 M * martin Anyone got a cris machine?
1118797226 M * martin :-)
1118797233 J * eXplasm2 explasm@p549F7CF2.dip.t-dialin.net
1118797249 Q * Snow-Man Remote host closed the connection
1118797332 M * martin I have to go but I'll keep persuing the hardware for the othr arch's and will assume NFS findings are of interest and wiki anything useful.
1118797348 M * Bertl great! have fun and TIA!
1118797375 Q * eXplasm Read error: Operation timed out
1118797404 M * martin Night
1118797407 Q * martin Quit: using sirc version 2.211+KSIRC/1.3.11
1118804034 J * badiane ~badiane@user-0cceriq.cable.mindspring.com
1118804050 M * Bertl welcome badiane!
1118804056 M * badiane hello 
1118804074 M * badiane I had to start a IRC client on my host 
1118804094 M * badiane I am having some problems with snat i guess 
1118804118 M * Bertl hmm?
1118804155 M * badiane the host can communicate with the web but the vserver just hangs
1118804167 M * Bertl maybe resolver issues?
1118804169 M * badiane I'm using shorewall 
1118804177 M * badiane could be let me check 
1118804179 M * Bertl maybe a wrong config ;)
1118804186 M * badiane that could also be 
1118804188 M * badiane :-)
1118804382 M * badiane the resolv.conf file is the same as the hosts
1118804386 M * badiane host's
1118804405 M * badiane but I've also tried to connect using ip 
1118804421 M * Bertl what about using tcpdump to check on the host?
1118804463 M * badiane for example i tried to connect to yahoo 66.94.234.13
1118804470 M * badiane hold on 
1118804513 M * badiane I've installed tcpdump 
1118804532 M * badiane what's a command set to use
1118804544 M * Bertl tcpdump -vvnei eth0 
1118804598 M * badiane tons of stuff going by very fast 
1118804625 M * Bertl well, maybe you should reduce the network traffic for a moment?
1118804639 M * Bertl (or look into the filter rules of the manpage)
1118804661 M * badiane ok 
1118805127 M * Bertl check out 'host' and 'proto tcp' or 'icmp' options
1118805510 J * Loki|muh_ loki@satanix.de
1118805510 Q * Loki|muh Read error: Connection reset by peer
1118805606 M * badiane I did a search with udp 
1118805617 M * badiane and from the vserver I pinged tekmd.com
1118805644 M * Bertl and? what did you see?
1118805665 M * badiane Iwhat I see is the ip of the vserver querrying the mindspring.com dns
1118805684 M * Bertl which probably isn't what you want?
1118805685 M * badiane all 3 of them 
1118805689 M * badiane nope 
1118805704 M * Bertl so it looks like your SNAT is wrong then=
1118805708 M * badiane it should be the host making the request on behalf of the vserver 
1118805708 M * Bertl s/=/?
1118805734 M * badiane before i had the host connected to my laptop which was serving as gateway 
1118805747 M * badiane that allowed me to use the gui to do other things
1118805767 M * badiane now the host is directly connected to the modem 
1118805768 M * Bertl how does your snat rule look like?
1118805780 Q * berni Ping timeout: 480 seconds
1118805817 M * badiane in the shorewall file I have eth0 as the outside interface and since my ip is dynamic I can't give it a 
1118805822 M * badiane specific ip to nat as 
1118805844 M * Bertl don't know shorewall rules, what is in the iptable chain?
1118805899 M * badiane i'm not sure let me look through the shorewall log first and then I'll do an iptable -L which might take a while 
1118805963 J * berni ~berni@svr01.mucip.net
1118805971 M * Bertl wb berni!
1118806104 M * badiane ok the logs say that whatever comes from the 192.168.1.8/29 subnet to 0.0.0.0 will be masqueraded
1118806126 M * Bertl that's fine, but there is nothing to masquerade from a vserver guest ;)
1118806146 M * Bertl you have to SNAT it (instead of masquerading)
1118806569 M * badiane hmmm
1118806569 M * badiane and if the ip is dynamic i can't quite give it an ip address to snat to 
1118806569 M * badiane @ least from the standpoint that I understand it 
1118806617 M * Aiken badiane I am on dialup, I use /etc/ppp/ip-up.local to redo my SNAT rules every time I connect to the net
1118806628 M * Aiken could you do something simillar?
1118806634 M * badiane sure 
1118806663 M * badiane but before I do I've just found a warning message in the shorewall log that I'm investigating
1118806780 M * Bertl masquerading happens on the 'forward' chain, when incoming packets get routed outbound
1118806794 M * Bertl the kernel has to 'map' ports in this case
1118806867 M * Bertl contrary to common misbelief, vserver guests do not have virtual interfaces (yet), so there is no routing except for the 'host' routing, which means an outbound connection does not use the forward chain at all
1118806909 M * Bertl thus it is not necessary/possible to map ports ... and masquerading will not apply
1118806954 M * Aiken Bertl you have said in plain text something that would have saved me a couple of days with google
1118806972 M * Bertl I say it every day just once ;)
1118807321 M * Bertl well, I guess I'm off to bed ... pretty tired actually ...
1118807359 M * Bertl guess my replies did sound a little harsh ...
1118807389 M * Bertl so sorry, it wasn't meant this way ... and I'm off to bed now ... cya tomorrow ...
1118807395 N * Bertl Bertl_zZ
1118807399 M * Aiken bye
1118809881 M * badiane I'm back 
1118809930 M * badiane no bertl it's not harsh it's just that the "concept" is not plainly spelled out in most of the vserver 
1118809934 M * badiane documentation out there 
1118809947 M * badiane aiken
1118809949 M * badiane are you there 
1118809966 M * Aiken I am here
1118809998 M * badiane I've place a little "clumsy" line in my shorewall script to extract the outside ip
1118810004 M * Aiken I worked it out from searching the vserver mailing lists and old usenet postings
1118810021 M * badiane when iver the the dhcp serves up a new address
1118810026 M * badiane and it work 
1118810030 M * badiane worked 
1118810038 M * Aiken PPPIP=`ifconfig ppp0 | grep 'inet addr' | awk -F: '{print $2}' | awk '{print $1}'`
1118810044 M * badiane yep 
1118810045 M * Aiken is it any worse than that?
1118810050 M * badiane same thing
1118810060 M * badiane except for the first awk I used cut
1118810108 M * badiane one of the conceptual problems I had and still have now that I have to rething everything 
1118810115 M * Aiken same effect :)
1118810119 M * badiane is that the vservers are not completely isolated 
1118810126 M * badiane or @ least very isolated 
1118810144 M * badiane so there is no difference between routing
1118810151 M * badiane it's all the same
1118810160 M * badiane I will be working with xen
1118810170 M * badiane I have an old k6 500MHz machine 
1118810187 M * badiane I wonder if vservers will run inside of a xen domain
1118810195 M * badiane that would be very, very interesting 
1118810202 M * Aiken that I don't know
1118810215 M * badiane virtualization inside paravirtualization
1118810222 M * Aiken I chose vserver instead of zen so I would not have to partition the memory
1118810232 M * badiane ahh yes 
1118810245 M * badiane they have set i up @ my job for staging servers 
1118810267 M * badiane but they can afford the dual machines with tons of mem
1118810423 M * Aiken more ram I might have looked at xen a bit longer but even with 1 gig of ram it does not take long to use it up
1118810578 M * Aiken plus I wanted something on a machine that was on 24x7 instead of this machine
1118810677 M * badiane I don't know if they're pushing them hard @ work but I'll find out what the perf is 
1118811561 M * Aiken had anyone had ntp running in a vserver?
1118811628 N * Doener|gone Doener
1118811636 M * Doener morning folks
1118811673 M * Doener badiane: someone once sent a patch to the ml that was supposed to make linux-vserver run under xen
1118812203 M * Aiken using http://linux-vserver.org/VirtualiseNTP for the final clue I have ntp 4.2.0 running in a vserver
1118813630 Q * Hollow Quit: /dev/zero
1118813780 J * Hollow ~Hollow@home.xnull.de
1118815291 Q * Aiken Read error: Connection reset by peer
1118815301 J * Aiken ~james@tooax6-182.dialup.optusnet.com.au
1118815659 Q * eyck Read error: Connection reset by peer
1118815926 J * eyck eyck@81.219.64.71
1118817060 Q * Hollow Read error: Operation timed out
1118817216 Q * erwan_taf Read error: Operation timed out
1118817245 J * erwan_taf ~erwan@81.80.43.67
1118817260 Q * TheSeer Read error: Operation timed out
1118817476 J * TheSeer ~theseer@border.office.salesemotion.net
1118817864 Q * Aiken Quit: Leaving
1118817893 Q * erwan_taf Read error: Operation timed out
1118817929 J * erwan_taf ~erwan@81.80.43.67
1118817937 J * erwan_ho ~erwan@konilope.dyndns.org
1118817980 Q * TheSeer Read error: Operation timed out
1118817991 J * TheSeer ~theseer@border.office.salesemotion.net
1118818686 Q * erwan_ho Remote host closed the connection
1118819069 Q * TheSeer Read error: Operation timed out
1118819081 J * TheSeer ~theseer@border.office.salesemotion.net
1118819715 Q * TheSeer helium.oftc.net charm.oftc.net
1118819715 Q * erwan_taf helium.oftc.net charm.oftc.net
1118819926 J * erwan_taf ~erwan@81.80.43.67
1118820683 J * TheSeer ~theseer@border.office.salesemotion.net
1118821372 J * serving- ~serving@217.164.247.199
1118821536 Q * serving Ping timeout: 480 seconds
1118822045 M * Pazzo anyone alive? one of my hosts is oopsing and I wonder if this could be vserver related...
1118822073 M * Pazzo (2.6.11.11-rc4)
1118822175 M * Pazzo http://vserver.gelf.net/oops.txt
1118822404 M * Pazzo the problem is caused by one folder in one single fileserver....
1118822465 M * Pazzo I can "stat /usr/share/doc/libmime-perl/html/Field/icons" but "ls -l /usr/share/doc/libmime-perl/html/Field/icons" fails with a Segmentation fault and this nice kernel oops
1118822563 M * Pazzo same thing also from host side: "ls -l /vservers/<vsname>/usr/share/doc/libmime-perl/html/Field/icons" -> segmentation fault
1118823100 J * prae ~prae@ezoffice.mandriva.com
1118823232 M * Doener good(?) morning Pazzo
1118823354 J * monrad ~monrad@213083190130.sonofon.dk
1118823783 M * Doener Pazzo: do you use xid tagging?
1118823968 J * KLFMANiK mann@people.uvt.tuke.sk
1118824013 Q * alexx Read error: Connection reset by peer
1118824042 M * Pazzo Doener: was doing so long time ago - but now surely not, partition is mounted without xid support
1118824078 J * alexx ~alexx@82.225.136.176
1118824118 M * Doener hmm... shouldn't matter anyways, i think...
1118824241 Q * KLFMANiK Quit: 
1118824443 M * Doener Pazzo: hm, is there a symlink involved?
1118824715 M * Doener nm
1118825283 J * martinh ~martinh@isa.voelcker.com
1118825360 M * martinh hello. i am trying to get an NXServer running inside a vserver. did anyone succeed in that?
1118825645 M * martinh i am able to login via nxclient. after a few seconds though (while kde is still starting), the client closes all of a sudden and without leaving a message!
1118825667 M * martinh i am using nxserver-freenx
1118825707 M * martinh which capabilities does a NXServer need??
1118825912 J * KLFMANiK mann@people.uvt.tuke.sk
1118826207 Q * martinh Quit: Verlassend
1118826346 M * KLFMANiK hello, where i can find info about Per Context Quota settings for latest stable 1.2.10 + kernel 2.4.30 + q0.14 ??? i found http://www.13thfloor.at/old/VServer/HowTo_LVMQ.shtml, http://www.13thfloor.at/old/VServer/HowTo_CTXQ.shtml, and some howtos for Disk Limits: http://linux-vserver.org/index.php?page=Quota+and+Disk+Limits, ... i'm using cq-tools from vallar.linuxfr.org/debian ... and now i have separate partition for guest, where i need quota ...
1118826350 J * martinh ~martinh@isa.voelcker.com
1118826625 Q * KLFMANiK Quit: [BX] Connection reset by panasync?
1118826805 M * Doener martinh: no messages? nothing in logs? did you try to strace the nxserver?
1118826896 J * KLFMANiK mann@people.uvt.tuke.sk
1118826937 M * Doener KLFMANiK: do you have a separate partition for _each_ guest, or shared by _all_ guests?
1118826951 M * Doener s/all/several/
1118826959 M * martinh seems like the logs don't work atm
1118826983 M * KLFMANiK Doener: now i'm testing per context quota on separate partition
1118827038 M * martinh Doener: nxserver doesn't write to my logfile although set up in node.conf!
1118827040 M * KLFMANiK Doener: maybe in future i'll use quota on shared partition (if no problems exists) ;-)
1118827055 M * Doener KLFMANiK: so there's one partition per guest, right?
1118827085 M * Doener then you don't need per context quota, but just a vroot device to allow quota ioctls from inside the vserver
1118827103 M * KLFMANiK Doener" right  .... i now testing howto http://www.13thfloor.at/old/VServer/HowTo_LVMQ.shtml ... but i don't know, if exists some newer info for 1.2.10 + q0.14
1118827160 M * martinh Doener: how can i strace nxserver, as it it started at session login?
1118827211 M * Doener martinh: i never used nxserver, but i guess it's called from some script, right? so you'd just have to modify that script then
1118827225 M * Doener KLFMANiK: hm, that one seems to be quite outdated
1118827264 M * Doener KLFMANiK: willing to add a page to the wiki? *g*
1118827276 M * KLFMANiK Doener: and it's not functional ... i can oversize quota for user ;-(
1118827326 M * Doener KLFMANiK: did you enable VROOT support in your kernel configuration?
1118827332 M * Doener (CONFIG_BLK_DEV_VROOT)
1118827367 M * martinh Doener: nxserver  is set as the login shell for the nx. that is the way it works. so i have to modify the login shell
1118827378 M * martinh Doener: i mean, /etc/passwd
1118827392 M * Doener ah, i see... 
1118827462 M * KLFMANiK Doener: yes ... CONFIG_BLK_DEV_VROOT=m
1118827477 M * Doener ok, do you have vrsetup?
1118827504 M * KLFMANiK yes, i have vrsetup
1118827559 M * KLFMANiK Doener: i feel progress in my problem ... a never used vrestup ;-)
1118827572 M * Doener ok, then first let's create a vroot device node
1118827587 M * KLFMANiK err ... vrsetup ... not vrestup ...
1118827608 M * Doener vrsetup /path/to/vserver/dev/hdv1 /dev/real-dev
1118827649 M * Doener real example: vrsetup /vservers/foo/dev/hdv1 /dev/vg1/vs-foo
1118827660 M * KLFMANiK i must stop this vserver now???
1118827665 M * Doener no
1118827678 M * Doener it just creates a device node
1118827739 M * Doener ok, now the vserver's mtab...
1118827803 M * KLFMANiK vrsetup /vservers/my_server/dev/hdv1 /dev/md8  .... ioctl(): Invalid argument
1118827819 M * Doener is the vroot module loaded?
1118827849 M * Doener the vserver's mtab should have a line like this: /dev/hdv1 / ufs rw,usrquota,grpquota 0 0
1118827889 M * KLFMANiK wait ... i must restart vserver guest ... ... mtab is ok
1118827895 Q * KLFMANiK Quit: [BX] Connection reset by panasync?
1118827973 J * KLFMANiK mann@people.uvt.tuke.sk
1118828018 M * Doener hm, why did you have to restart the vserver?
1118828034 M * KLFMANiK Doener: now i get: ioctl(): Not a typewriter
1118828077 M * Doener from vrsetup?
1118828104 M * KLFMANiK yes
1118828157 M * Doener ah, sorry, we didn't create the device node
1118828191 M * Doener mknod /path/to/vserver/dev/hdv1 b 4 0
1118828193 M * KLFMANiK yes .. it's only simple file
1118828213 M * Pazzo re
1118828225 M * Pazzo thnx Doener, had to leave before
1118828232 M * KLFMANiK ok, mknod done, vrestup done
1118828248 M * KLFMANiK vrsetup done ...
1118828248 M * Pazzo moved all vservers to another host, no vserver running, nothing (but info) in /proc/virtual
1118828260 M * Pazzo lsof | grep vse => nothing
1118828269 M * Pazzo umount /vservers => busy
1118828274 M * Pazzo ??
1118828305 M * Pazzo reboot and fsck?
1118828307 M * KLFMANiK Pazzo: try mount
1118828328 M * KLFMANiK Pazzo: maybe some proc in vserver remains mounted
1118828342 M * Doener KLFMANiK: guess that won't help... it's probably in another namespace...
1118828369 M * Doener Pazzo: try with: chcontext --ctx 1 grep vse /proc/*/mounts
1118828401 M * KLFMANiK Doener: ok ;-)
1118828402 M * Pazzo hmm... uptime is 3 days - kernel 2.6.11.11-vs2.0-rc4, utils 0.30.207-7 from sid => can there still be issues with stopping vservers??
1118828463 M * Doener even if there were, there shouldn't be some mounts hanging around in no mans land, that would be quite a new issue
1118828486 M * Doener KLFMANiK: ok, are you using the new or the old configuration scheme?
1118828501 M * Doener (old = one file, new = directory)
1118828510 M * KLFMANiK Doener: old ... but with new utils from Sarge
1118828539 M * Doener ok, then add CAP_QUOTACTL to the vserver's capabilites, IIRC that was called S_CAPS or sth. like that
1118828565 M * Doener and this time you have to restart the vserver ;)
1118828588 M * KLFMANiK ok ... 
1118828608 Q * KLFMANiK Quit: ... i must logout, because i'm testing quota on this vserver ;-))
1118828709 M * Pazzo Doener: there are many /proc/<pid>/mounts:/dev/md1 /vservers ext3 rw 0 0
1118828731 M * Pazzo (and proc/self of course)
1118828733 M * Pazzo nothing else
1118828755 J * KLFMANiK mann@people.uvt.tuke.sk
1118828792 M * Doener maybe the oopses created some dangling half-open files or so...
1118828855 M * KLFMANiK Doener: THANKS ... it works ;-)
1118828860 M * Pazzo ok, I'll reboot and then fsck the /vservers partition - thnx in the meantime
1118828891 M * Doener KLFMANiK: great :) would you care about adding some page to the wiki? (see topic ;)
1118828917 M * KLFMANiK Doener: ok .. i try
1118828923 M * Doener thanks!
1118828944 M * KLFMANiK Doener: and what about shared partitions and context quota + disk limits???
1118828968 M * Doener never used those... 
1118828998 M * Pazzo ???? reboot ok, fsck ok - files ok
1118829019 M * KLFMANiK ok ... thanks anyway
1118829031 M * Pazzo btw: while partition was mounted "fsck -n /dev/md1" told me that there have been errors
1118829040 J * rs ~rs@staff.lycos.fr
1118829102 M * Pazzo Free blocks count wrong (31585637, counted=33087062).
1118829113 M * Pazzo Free inodes count wrong (18001879, counted=18321137).
1118829119 M * Pazzo (nothing else)
1118829188 M * Pazzo (trying to run fsck -c now)
1118829256 P * erwan_taf Leaving
1118829960 Q * KLFMANiK Quit: thanks Doener ...
1118830178 M * fwiffo moin
1118830232 M * DaPhreak morning fwiffo 
1118830239 M * fwiffo uh hi
1118830253 M * fwiffo btw, the patch is not suitable for rbac...
1118830257 J * yarihm ~yarihm@217-162-113-200.dclient.hispeed.ch
1118830265 M * fwiffo working on a fix for this
1118830318 M * fwiffo it's just a bit difficult to debug a grsec kernel with all randomizing options enabled :-D
1118830388 M * DaPhreak yeah ;)
1118830500 M * pusling anything vserver related I should be aware of when upgrading kernel from 2.4.x to 2.6.x ?
1118830543 M * fwiffo hm, if you use the same vserver version, then no ;-)
1118830561 M * Doener fwiffo: hm, that would be impossible...
1118830582 M * fwiffo hullo Doener
1118830615 M * fwiffo no forwardports? ;-)
1118830644 M * Doener pusling: http://linux-vserver.org/alpha+util-vserver (last paragraph) has some information... apart from that it should just work
1118830652 J * KLFMANiK mann@people.uvt.tuke.sk
1118830682 M * Doener fwiffo: the 2.4 stuff is considered legacy by now, so why should we? ;)
1118830708 M * fwiffo just for fun ;-)
1118830827 M * KLFMANiK Doener: now i'm trying to use quota on two vservers on two separate partitions ... on first it works, but on second: vrsetup exits with: ioctl(): Device or resource busy ... i stopped vserver and vrsetup exits with same error ... i tried mknod /dev/hdu1 but withou success
1118830848 M * fwiffo hm, any idea how to use crash on grsec with randomized kernel mem?
1118830872 M * Doener KLFMANiK: did you use an other minor number?
1118830885 M * pusling hm... where do I find 'namespaces' in make menuconfig ?
1118830897 M * KLFMANiK Doener: .... uff no, sorry ;-) my mistake
1118830920 M * Doener pusling: nowhere ;) that's always built-in
1118830941 N * Loki|muh_ Loki|muh
1118830943 M * pusling Doener: ahh.. ;) That was why I couldn't find it ;)
1118830950 M * Doener if your vservers use namespaces is chosen in the vserver's configuration...
1118830970 M * Doener legacy: never use namespaces. new style: use them by default, you can choose not to use them
1118830993 M * pusling my vservres are made legacy, but later converted to new style ;)
1118831003 M * pusling is flowerpage the right place to look_?
1118831033 M * Doener for the namespace-related options in the new style configuration? yes
1118831201 M * pusling Did I understand this the right way: Not having the file /etc/vservers/.defaults/nonamespace enables namespace use ?
1118831304 M * fwiffo uh...
1118831351 M * fwiffo the vserver syscall with the new_s_context cmd is only included with legacy option enabled O_o
1118831400 M * fwiffo but vserver-info uses that, does it mean i use ald tools O_o oh, wait, checking that...
1118831461 M * KLFMANiK Doener: quotacheck: error (2133571364) while opening /dev/hdu1
1118831495 Q * dsoul Ping timeout: 480 seconds
1118831510 M * Doener pusling: neither that file, nor a nonamespace file in the vserver's configuration itself
1118831573 M * pusling okay
1118831597 M * Doener KLFMANiK: and /dev/hdu1 is there?
1118831612 M * fwiffo ah, okay, that's replaced with vc_create_context()
1118831639 M * Doener btw, why hdu1 and not hdv1?
1118831649 M * KLFMANiK Doener: yes ... 
1118831669 M * KLFMANiK Doener: because it's another partition with another vserver
1118831714 M * KLFMANiK Doener: brw-r--r--  1 root root 4, 2 Jun 15 12:33 /dev/hdu1
1118831727 M * Doener well, each vserver got its own device node, right? and they're in different directories, so you could call them all hdv1... but shouldn't matter, just been curious ;)
1118831768 M * Doener what does the vserver's mtab contain?
1118831786 M * KLFMANiK Doener: yes ...i know ... but it's for my simplification ... another vserver, another partition, another hdX
1118831830 M * KLFMANiK i made vrsetup /var/lib/vservers/post/dev/hdu1 /dev/hda8 before ... mtab: /dev/hdu1 / ext3 rw,usrquota,grpquota 0 0
1118831921 M * Doener change ext3 to ufs
1118832031 M * KLFMANiK BINGO ...it works ..... but first vserver has ext3 in mtab ...
1118832062 M * KLFMANiK BTW - only ext2/ext3 is supported for this type of quota?
1118832101 M * Doener that should work with any fs AFAIK
1118832137 M * Doener ( as long as the fs supports quota... is there any that doesn't? )
1118832149 M * Doener time for lunch... back in a few...
1118832152 N * Doener Doener|gone
1118832306 J * Doener` ~doener@p54874DCB.dip.t-dialin.net
1118832740 Q * Doener|gone Ping timeout: 480 seconds
1118832894 N * Doener` Doener
1118832903 M * Doener back now...
1118832908 M * fwiffo wb doener
1118833631 Q * case Ping timeout: 480 seconds
1118833636 Q * Pazzo Ping timeout: 480 seconds
1118833720 Q * KLFMANiK Quit: BitchX: no additives or preservatives
1118833843 Q * rs Quit: rs
1118834113 J * case ~case@donpanic.faveve.uni-stuttgart.de
1118834152 J * Pazzo ~Pazzo@host130-250.pool8172.interbusiness.it
1118835747 J * ArnY ~arny@copernic.lina.sciences.univ-nantes.fr
1118835839 M * ArnY hello, i'm migrating vservers from the stable branch (2.4 kernel) to a 2.6 host server.. when i start my new vservers, it doesn't look like any arp packet are sent to adverstive the new MAC of the vserver...
1118835857 M * ArnY is that a normal behavior?
1118836027 J * IceTi 125@mw635449.rz.ruhr-uni-bochum.de
1118836029 M * IceTi hi
1118836116 J * rs ~rs@staff.lycos.fr
1118836148 M * ArnY Hiya IceTi & Rs
1118836161 M * Pazzo hmm... how can I get detailed information about bad blocks??
1118836194 M * ArnY with badblocks ?
1118836197 M * ArnY "badblocks - search a device for bad blocks"
1118836299 M * Pazzo thnx - alternative? fsck wasn't very informative (maybe I used wrong params!?)
1118836379 M * Pazzo e2fsck -c /dev/md1
1118836414 M * fwiffo Pazzo: maybe you have to use a destructive test...
1118836420 M * ArnY well, i couldn't say.. we don't really bother keeping harddrives with badblock here
1118836421 M * Pazzo hmmm... is doing the same check as "fsck -c /dev/md1" already did... let's hope it will be more informative
1118836454 M * Pazzo me too - but I want to find out IF there are bad blocks
1118836472 M * Pazzo had really strange oopses today :(
1118836485 M * fwiffo badblocks -o output_file -w /dev/md1
1118836497 M * fwiffo but use anly if you don't have to loose anything
1118836513 M * fwiffo s/anl/onl/
1118836614 M * Doener ArnY: the kernel sends an arp reply when asked for, not when the address is created
1118836773 M * ArnY i thought the vserver startup script was advertizing the new mac when starting the vserver?
1118836782 A * ArnY must have misread some doc there
1118836940 M * Doener there's no new mac...
1118836963 M * Doener it's the box with the same nic, thus the same mac ;)
1118837005 M * ArnY yeah, but the new mac for the old ip, that's what i meant
1118837037 M * Doener hm? new ip for old mac you mean?
1118837214 M * ArnY i migrated a vserver from a server to another one
1118837245 M * ArnY it kept its ip, but since the server changed, the mac changed too
1118837251 M * ArnY same ip, new mac
1118837361 Q * Loki|muh Ping timeout: 480 seconds
1118837441 M * Doener ah, now i see :)
1118837465 J * Loki|muh loki@satanix.de
1118837665 M * ArnY for some reason it takes a while for the new vserver to pick up the packets
1118837681 M * ArnY i guess it's due to our crappy alcatel router
1118837802 M * Doener you could sent some unsolicited arp reply with arping
1118837815 M * ArnY arping.. ah.. might be what i need indeed
1118838749 J * Aiken ~james@tooax6-159.dialup.optusnet.com.au
1118839577 M * Pazzo "Checking for bad blocks (read-only test): done                        112"
1118839595 M * Pazzo what does this 112 mean? 112 bad blocks?
1118839620 M * fwiffo look into the output file
1118839633 M * fwiffo all blocks which are listed there are bad blocks
1118839668 M * fwiffo but you told you got weird kernel oops...
1118839675 M * fwiffo try doing a memtest
1118839717 M * fwiffo that's mostly the cause for strange bohaviour of programs...
1118839810 M * fwiffo s/ly/ likely/
1118839879 M * Pazzo hehe... you're right :-)
1118839885 M * Pazzo ehm... :-(
1118839901 M * Pazzo host is 400 miles away :(
1118839902 N * Bertl_zZ Bertl
1118839906 M * Doener morning Bertl 
1118839906 M * fwiffo hehehe
1118839910 M * fwiffo moin bertl
1118839913 M * Pazzo moin Bertl
1118839915 M * Bertl morning folks!
1118839926 M * fwiffo Pazzo: had the same problem some weeks ago
1118840003 Q * ruuth Quit: Nettalk6 der Freeware IRC-Client
1118840231 J * miller7 ~none@adsl16-49static.access.acn.gr
1118840240 M * miller7 hello guys
1118840245 M * fwiffo hi miller
1118840284 M * miller7 :)
1118840306 M * miller7 can someone tell me why my box keeps an ip of a vserver which is not started?
1118840316 M * miller7 it still has the virtual interface on
1118840320 M * miller7 is this normal?
1118840322 M * ArnY dab shutdown script?
1118840347 M * ArnY err bad
1118840351 M * miller7 hmmm
1118840359 M * miller7 how should it release the IP?
1118840363 M * miller7 can I release it manually?
1118840395 M * fwiffo ifconfig iface down ?
1118840424 M * fwiffo or what do you mean with virtual interface exactly
1118840425 M * fwiffo ?
1118840429 M * fwiffo ngnet?
1118840479 M * miller7 the interface that vserver script creates
1118840498 J * Val ~val@v41.ath.cx
1118840508 M * Val Hi
1118840515 M * miller7 oh well... I have to check my scripts again :(
1118840523 M * miller7 probably I've done somthing stupid there
1118840527 M * miller7 anyway, thanks guys
1118840529 M * fwiffo np
1118840548 M * fwiffo moin val
1118840635 P * miller7 
1118840880 M * Doener Bertl: is this really necessary? 
1118840881 M * Doener memset (new, 0, sizeof(struct vx_info));
1118840886 M * Doener atomic_set(&new->vx_usecnt, 0);
1118840907 M * Bertl the atmonic_set?
1118840907 M * Doener (left out some code, but that doesn't touch new->vx_usecnt)
1118840910 M * Doener yep
1118840930 M * Doener it's all set to 0 anyways, isn't it?
1118840931 M * Bertl well, atomics are 'special' i.e. each arch can implement them as they 'like'
1118840953 M * Bertl they have no guarantee that all bits are available, and they have no guarantee that this is true for all archs
1118841014 M * Bertl so while this is probably true on all current archs, you can not count on it in general
1118841026 M * Doener ok... strange things happen... ;)
1118841040 M * Doener new->vx_state = 0; what about this?
1118841067 M * Doener btw, I'm trying to add a slab cache atm... thus i stumbled upon that stuff ;)
1118841068 M * Bertl this is gratious initialization, I'd say ...
1118841084 M * Bertl with a slab, you have to do that ...
1118841093 M * Bertl (sidenote)
1118841130 M * Doener yep... thought about moving some stuff into a constructor, but there's nothing to move ;)
1118841163 M * Bertl hmm ...
1118841163 M * Doener just wondering about the current things... but you did that in preparation of a slab, right? ;)
1118841202 M * Bertl yeah, well, the slab is planned for some time ... so although I did not explicitely look into that direction, it probably influenced the design a little
1118841567 J * id ~id@relax-media.softwarezentrum.de
1118841663 M * id gruetzi #vserver
1118841672 M * fwiffo hi id
1118841781 M * ArnY hum.. any idea about this error in a vserver: "Error opening file /proc/kmsg for reading (Operation not permitted)" ?
1118841834 M * Bertl inside a guest?
1118841839 M * ArnY yes
1118841852 M * Bertl well, you are not allowed (Operation not permitted) to do that inside a guest
1118841868 M * Bertl and for a good reason, because you could read sensitive kernel information from tere
1118841871 M * fwiffo wuahaha *cry*
1118841871 M * Bertl *there even
1118841885 M * fwiffo i get eperm since weeks...
1118841895 M * fwiffo s/rm/&s/
1118841931 M * Bertl hmm, so?
1118841985 M * fwiffo never mind, that's just because i cannot read "Operation not permitted" anymore without goin' insane
1118841995 M * fwiffo :-D
1118842030 Q * IceTi Ping timeout: 480 seconds
1118842225 M * ArnY dang! the recent syslog-ng versions replaces klogd and tries to access /dev/kmsg 
1118842226 M * ArnY grrr
1118842247 M * ArnY we used disable klogd and leave syslog-ng
1118842309 M * fwiffo hm, mom...
1118842313 M * fwiffo i'll try that...
1118842381 M * fwiffo (because i wanted to use syslog-ng inside the guests, too)
1118842409 M * ArnY well, you can disable the kernel source driver in the config file
1118842414 M * ArnY i just found this out
1118842435 M * ArnY there's a new source s_sys { pipe ("/proc/kmsg" log_prefix("kernel: ")); 
1118842444 M * ArnY which obviously causes problems
1118842464 M * fwiffo hrm, usually you don't have to read out kernel messages inside a guest...
1118842477 M * eyck   just remove it
1118842484 A * ArnY nods
1118842491 M * fwiffo imho that's a thing which should be done on the host
1118842498 M * eyck this is what we've been doing for years ;) 
1118842548 M * ArnY eyck: i installed syslog-nng for the first time in a vserver (the previous install was done by a coworker who forgot to leave a note about that problem)
1118842591 M * eyck there are few other things you might like to change in guests, 
1118842633 M * eyck  debian-newvserver used to fixed those, but now it's deprecated in favor of new vserver sth build,
1118842650 M * eyck which unfortunatelly leaves your guest half-done to say the least..
1118842680 A * ArnY had to had a lot of stuff in initpost for his fc3 vservers
1118842684 M * ArnY s/had/add
1118842737 M * ArnY the vserver-build util should copy /etc/localtime from the host to the guest system.. 
1118842971 Q * Aiken Ping timeout: 480 seconds
1118843042 M * Bertl ArnY: /dev/kmsg is not an issue
1118843064 M * ArnY no, it's not
1118843091 M * Bertl what kernel version do you use?
1118843105 M * Bertl I misread the kmsg as kmem ... somehow
1118843107 M * ArnY i just asked if it could be bypassed with a vserver setting.. but it's not big deal, i'lljust disable the kernel source freom syslog-ng's config file
1118843132 M * ArnY 2.6.11.5-vs1.9.5
1118843141 M * Bertl the 2.0-* branch should support a virtual kmsg ...
1118843151 M * Bertl so it should work out of the box there ...
1118843179 M * ArnY hum.. any special settings in the vserver config to enable this?
1118843185 M * Bertl so updating to the latest release candidate should do it
1118843218 M * ArnY ah.. a little too late for me ;-)
1118843409 M * ArnY arg
1118843422 M * Bertl well, you probably should update anyway at least when the stable release is out
1118843439 M * ArnY stopping a vserver running on eth1.34 just shutdown the interface.. i had another vserver running on this interface.. how could that happen
1118843450 M * Bertl Doener: I found an interesting bug ...
1118843471 M * Bertl ArnY: that is default linux behaviour if primary/secondary is assigned in an unfortunate way
1118843502 M * Bertl Doener: try: 'vcontext --create --xid 42 -- sleep 100 &'
1118843512 M * Bertl then 'vattribute --xid 42 --flag ~^32'
1118843519 M * Doener will it crash the box? i.e. should i use qemu?
1118843531 M * Bertl yes and no, be careful ...
1118843689 M * ArnY is there a way know if a vlan interface is in use by a vserver before removing it?
1118843738 M * ArnY +to
1118843740 M * Doener hm, nice
1118844511 M * Pazzo oops ;-)
1118844594 M * ArnY Doener: arping does a great job, thanks
1118844605 M * Doener yw
1118844635 M * Bertl Doener: the basic issue is that the context lookup in setup state is not supposed to succeed ...
1118844655 M * Bertl (so it should be easy to fix)
1118844659 M * Pazzo   ~^32 => VXF_STATE_SETUP ? what's that?
1118844671 M * Bertl Pazzo: better forget about it ;)
1118844690 M * Bertl okay, leaving now ... back later ...
1118844697 M * Pazzo cu Bertl ;)
1118844703 N * Bertl Bertl_oO
1118844895 M * Doener Bertl_oO: how is the setup state useful then? it's there, but you can't use it?
1118845738 P * yarihm Leaving
1118846115 M * Pazzo Doener: what is VXF_STATE_SETUP used for?
1118846130 M * Pazzo or should/should not be used for
1118846185 M * Doener internal flag to determine whether a context is in setup state...
1118846224 M * Pazzo btw: anyone using the libata-dev patchset? -> http://www.kernel.org/pub/linux/kernel/people/jgarzik/libata/
1118846301 M * Pazzo Doener: is it sane to launch "vserver xy stop" immediately after running "vserver xy start" without waiting for start being finished?
1118846306 J * romke ~romke@procyon.romke.net
1118846338 M * Doener it should be, as there are some locks in the tools...
1118846348 M * Doener from the kernel side, it shouldn't matter at all
1118846564 Q * locksy Ping timeout: 480 seconds
1118846625 M * Pazzo thnx Doenr
1118847604 Q * romke Quit: [resboot]
1118847830 N * BobR_oO BobR
1118848135 N * BobR BobR_oO
1118848244 M * ArnY how does a vserver bind an interface now it doesn't use the aliases?
1118848329 M * Doener it never bound to an interface, binding is based on ip addresses
1118848367 M * ArnY i mean, it used to create an alias of the network interface per vserver
1118848405 M * ArnY how was this replaced?
1118848485 M * Doener actually, there's no alias for the _interface_, but the interface _addresses_ get a name. from the kernel's point of view, no such name is necessary (at least on 2.4+)
1118848523 M * Doener the old ioctl interface relied on interface address names, thus ifconfig can only show named interface addresses
1118848547 M * ArnY hum
1118848550 J * romke ~romke@procyon.romke.net
1118848562 M * Doener since 2.4 (IIRC) there's the netlink interface, which doesn't need named interface addresses
1118848598 M * Doener you _can_ still assign a name, but you don't _have_ to
1118848621 M * Doener the iproute tools use the new netlink interface and are thus able to display such nameless interface addresses
1118848670 M * ArnY ah.. i see
1118848680 M * ArnY ip addr show them
1118848685 M * Doener the old tools always created named interface addresses, because the used ifconfig
1118848714 M * Doener the new tools, by default, do not assign names to the interface addresses, but you can choose to
1118848728 M * ArnY how would i do that?
1118848756 M * Doener IIRC you have to create a file called "name" in the interface configuration directory, take a look at the flower page for details
1118848774 M * ArnY ah, that's the name setting
1118848782 M * ArnY i wasn't sure what it was for
1118848783 M * ArnY thanks
1118848787 M * Doener you're welcome
1118849398 N * ArnY ArnY[AwaY]
1118851384 Q * martinh Quit: Verlassend
1118851452 Q * rs Quit: rs
1118852255 Q * Val Read error: Connection reset by peer
1118853861 J * fwiffo_ ~mmrnmrhm@p5499D85A.dip.t-dialin.net
1118854154 M * pusling I get this on my newly compiled kernel when starting vservers: "idr_remove called for id=1 which is not allocated" and then a about 20 lines large dump of [<somecode>] description
1118854241 Q * fwiffo Ping timeout: 481 seconds
1118854259 M * pusling (but stuff seems to work)
1118854457 M * pusling this is the dump repeated many times http://pastebin.com/299949 (but I don't know if it is vserver related or just happens while starting a vserver) 
1118854777 J * rs ~rs@Laubervilliers-151-13-4-57.w82-127.abo.wanadoo.fr
1118854838 J * Hollow ~Hollow@home.xnull.de
1118855210 Q * rs Read error: Connection reset by peer
1118855369 J * rs ~rs@Laubervilliers-151-13-4-57.w82-127.abo.wanadoo.fr
1118856044 Q * dotwaffle Quit: leaving
1118856502 Q * prae Quit: Execute Order 69 !
1118856737 J * locksy ~locksy@mrtg.sisgroup.com.au
1118857710 Q * locksy Ping timeout: 480 seconds
1118858310 N * fwiffo_ fwiffo
1118858350 Q * rs Quit: rs
1118860951 J * Mauritz guest@200.103.121.80
1118860961 P * Mauritz 
1118861184 M * pattieja_ fwiffo, hello
1118861325 N * pattieja_ pattieja
1118861495 M * pattieja after making those changes to successfully get Postfix operational on the IP addresses of eth0 and not binding to 0.0.0.0, the Mailman instance no longer functions properly
1118861519 M * pattieja mail sent to a mailing list is received by the list and archived, but no users receive e-mail messages sent to the list
1118861699 M * fwiffo hullo
1118861757 M * fwiffo what does the logfile tell you?
1118862061 M * pattieja fwiffo, which log file?
1118862070 M * fwiffo the postfix one
1118862082 M * pattieja I watched mail.log as I sent a message and it does not appear to contain anything out of the ordinary
1118862111 M * pattieja it passes the message to /var/lib/mailman/mail/mailman post <listname>
1118862127 M * pattieja that then, apparently, runs the message through spamd
1118862152 M * fwiffo O_o
1118862184 M * fwiffo shouldn't spand be handled by postfix?
1118862210 M * fwiffo hm, i think that's off topic
1118862221 M * fwiffo ... switch to query
1118862427 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il
1118863329 Q * pusling Quit: leaving
1118863345 J * pusling ~pusling@195.215.29.124
1118864429 J * locksy ~locksy@mrtg.sisgroup.com.au
1118864629 J * shuri shuri@64.235.209.226
1118864813 M * shuri Hi
1118864820 M * fwiffo 'lo
1118864980 M * Doener evening shuri 
1118865566 M * shuri Thx Doener 
1118865627 J * erwan_ho ~erwan@konilope.dyndns.org
1118866081 Q * monrad Quit: Leaving
1118866335 Q * eXplasm2 Remote host closed the connection
1118866713 J * eXplasm explasm@p549F7CF2.dip.t-dialin.net
1118869101 Q * eXplasm Remote host closed the connection
1118869752 J * hansi33 ~hansi33@193.80.108.229
1118869823 M * hansi33 inside my vserver i get this message from mysql
1118869825 M * hansi33 setrlimit couldn't increase number of open files to more than 1024 (request: 2258)
1118869837 M * hansi33 does anyone have an idea what to do?
1118869974 Q * erwan_ho Remote host closed the connection
1118870013 J * rs ~rs@imhotep.rhapsodyk.net
1118870359 J * yarihm ~yarihm@217-162-112-73.dclient.hispeed.ch
1118870551 Q * hansi33 Quit: ChatZilla 0.9.61 [Mozilla rv:1.7.7/20050507]
1118870571 J * hansi33 ~hansi33@193.80.108.229
1118870723 J * brc bruce@200141099215.user.veloxzone.com.br
1118870728 M * brc is quota inside a vserver possible ?
1118870777 M * pattieja brc, I believe so, but I haven't tried it yet
1118870787 M * pattieja there's specifically a patch that allows for it
1118870940 M * Doener brc: what kind of quota?
1118870964 M * Doener on a shared partition, or on a partition that is used only for that vserver?
1118871029 M * Doener I know the latter one works, don't know about the former one with 2.6, but it is possible with 2.4
1118871056 M * brc Doener: shared partition
1118871066 M * brc well
1118871071 M * brc sine we cant use quota, do you know of any other good solution ?
1118871082 M * brc so that a web user who can upload wont fill up all the vserver's virtual hard disk ?
1118871191 M * Doener no...
1118871224 M * pusling a daily 'du -hs /vservers | mail sysadmin' ;)
1118871230 M * Doener ask Bertl if shared partition quota is possible with 2.6 by now
1118871357 M * brc :(
1118871412 M * pusling I have from outside mounted a partition into the webserver. It works fine, but writing df inside the only show the / partition
1118871595 M * brc the problem is that therewill be several vservers
1118871617 M * brc and each of them might will have its own webserver :)
1118872237 J * darksoul darksoul@pingu.ii.uj.edu.pl
1118872735 J * ruuth VooDoo@topas.informatik.uni-ulm.de
1118873444 J * Aiken ~james@tooax6-217.dialup.optusnet.com.au
1118873526 Q * gregster Remote host closed the connection
1118873527 J * gregster ~gregor@greart.de
1118873960 Q * badiane Ping timeout: 480 seconds
1118874240 J * badiane ~badiane@user-0cceriq.cable.mindspring.com
1118874540 Q * ruuth Read error: Connection reset by peer
1118874640 Q * hansi33 Ping timeout: 480 seconds
1118874937 J * ciphernaut ~a@61.88.18.130
1118874947 M * ciphernaut morning all
1118875678 M * mugwump hi ciphernaut 
1118875983 Q * rs Quit: rs
1118875998 J * rs ~rs@imhotep.rhapsodyk.net
1118876122 P * romke 
1118876172 J * hwarrier hwarrier@adsl-67-112-26-96.dsl.snfc21.pacbell.net
1118876656 Q * hwarrier Ping timeout: 480 seconds
1118876740 Q * Zoiah Ping timeout: 480 seconds
1118877196 Q * yarihm Quit: Leaving
1118879042 N * Bertl_oO Bertl
1118879046 M * Bertl evening folks!
1118879141 M * Doener evening Bertl
1118879151 M * Aiken good morning
1118879299 M * Bertl hey Doener! Aiken!
1118879413 M * Doener still working on your fundings?
1118879462 M * Doener if so, same here... :/
1118879630 M * Bertl hmm ... well ...
1118879645 M * Doener btw, there's an article that covers solaris' zones (at least to some extent) in the german Linux Magazin
1118879669 M * Bertl is it online somewhere?
1118879675 M * Doener just checking
1118879741 M * Doener no :(
1118879797 M * Doener but it's from the user's perspective anyways... basically the interesting information seems to be:
1118879828 M * Doener they got not separation between context 0/1, there's just a global zone that sees everything, and the zones themselves
1118879907 M * Bertl okay
1118879911 M * Doener filesystem virtualisation seems to go a step further... "Direct access to device nodes has been limited and a zone only sees its own sector of the filesystem (cf. chroot)."
1118879923 M * Doener whatever a "sector" may be in the solaris world
1118879938 M * Bertl a chroot/namespace, I'd say
1118879956 M * Bertl so the device node blocker is a feature IMHO
1118879974 M * Doener oh, and they got ressource guarantees
1118879986 M * Bertl hmm, interesting ...