1105488397 Q * nayco Quit: Bonne nuit !
1105488725 M * chrish01 well, my asp.net stuff works pretty well...but im not gonna post any patches...just wait till lluis' stuff is in and rewrite it
1105491290 J * dsanta ~santa@c68.190.156.105.roc.mn.charter.com
1105491384 N * Doener_zZz Doener
1105491392 M * Doener evening!
1105491803 M * albeiro evening Doener :]
1105504193 Q * lilo Quit: leaving
1105505107 J * nox- ~nox@c207113.adsl.hansenet.de
1105505455 Q * nox Ping timeout: 480 seconds
1105505467 N * nox- nox
1105506286 J * lilo ~lilo@lilo.usercloak.oftc.net
1105506902 T * * http://linux-vserver.org/ | latest stable 1.29, devel 1.3.9, 1.9.3, ng8.7
1105506902 T * Doener -
1105506903 J * eyck eyck@81.219.64.71
1105506904 J * mugwump ~samv@210-54-92-184.ipnets.xtra.co.nz
1105506905 J * tchan_ ~tchan@c-24-13-81-164.client.comcast.net
1105506909 J * sannes ~ace@home.skarby.no
1105506914 J * ensc ~ircensc@ultra.csn.tu-chemnitz.de
1105506918 J * locksy ~locksy@mrtg.sisgroup.com.au
1105506920 J * grecea ~grecea@h-195-22-237-74.mdl.net
1105506926 J * Seraph kk@212.9.177.15
1105506927 J * no_maam ~erik@194.95.226.17
1105506936 J * TheSeer ~theseer@212.12.45.62
1105507229 J * sladen paul@80.1.73.116
1105508578 J * Loki|muh_ loki@satanix.de
1105508657 Q * ola uranium.oftc.net arion.oftc.net
1105508657 Q * Loki|muh uranium.oftc.net arion.oftc.net
1105508726 J * ola ~ola@c-adt-5.ataco.se
1105508741 Q * ola Remote host closed the connection
1105508742 J * ola ~ola@c-adt-5.ataco.se
1105513940 Q * are|afk Ping timeout: 480 seconds
1105515111 Q * flock Ping timeout: 480 seconds
1105519238 T * * http://linux-vserver.org/ | latest stable 1.29, devel 1.3.9, 1.9.3, ng8.7
1105519238 T * Doener -
1105522043 J * rs rs@ice.aspic.com
1105522069 M * rs hi folks
1105522294 M * Loki|muh_ good morning rs
1105522689 Q * BWare Quit: using sirc version 2.211+KSIRC/1.3.10
1105523699 J * tchan ~tchan@c-24-13-81-164.client.comcast.net
1105523800 J * serving_ ~serving@213.186.185.143
1105523800 Q * serving Read error: Connection reset by peer
1105523815 Q * tchan_ Ping timeout: 480 seconds
1105524786 J * pulsar ~pulsar@82.209.246.234
1105524935 N * chrish01 chris|sleep
1105525720 Q * grecea Remote host closed the connection
1105525899 N * Bertl_oO Bertl
1105525911 M * Bertl morning folks!
1105525973 J * grecea ~grecea@h-195-22-237-74.mdl.net
1105526038 M * Bertl welcome grecea!
1105526172 M * Loki|muh_ moin Bertl 
1105526176 N * Loki|muh_ Loki|muh
1105526188 M * ndim Morning everyone. What was the status of NGNET and IPv6? NGNET without userspace tools to set them up, IPv6 not implemented, but easy to add?
1105526263 J * BWare ~bware@212.26.196.41
1105526553 M * Bertl ndim: well, first no, tools are all there, second yep
1105527014 M * ndim Bertl: OK. Then I'm going to give NGNET a try in preparation for IPv6 support. I have a local webserver running on IPv6 only (I'm sick and tired of port forwarding, and it is pretty neat to have access to my internal servers from the outside when required), and I'd like to move that into a vserver. I can't say anything about the timeframe to do that yet, though.
1105527042 M * Loki|muh Bertl: you said that with the the vserver-patch there is no need for enabling HIGHMEM... but dmesg says the following to me:
1105527045 M * Loki|muh |>Warning only 896MB will be used.
1105527047 M * Loki|muh |>Use a HIGHMEM enabled kernel.
1105527050 M * Loki|muh |>896MB LOWMEM available.
1105527054 M * Loki|muh 2.6.9-vs1.9.3
1105527078 M * Bertl get 2.6.10-vs1.9.3.14+ and enable a 2/2 split with 2GB memory
1105527121 J * amit ~Amit@220.225.32.98
1105527130 M * Bertl welcome amit!
1105527140 M * amit hello Bertl 
1105527193 M * Loki|muh ah! :) which version should I use for production? is there a changelog for 1.9.3.xx?
1105527243 M * Bertl sometimes ... (actually it's 1.9.3.11+) http://linux-vserver.org/ChangeLog26
1105527507 M * pulsar Where I cat get documentation about "Hard CPU scheduler" ?
1105527514 M * Loki|muh and which version is considered stable enough for production?
1105527548 M * Bertl pulsar: http://linux-vserver.org/Linux-VServer-Paper-06
1105527563 M * Bertl Loki|muh: 1.2x, therefore it's called 'stable' branch ;)
1105527585 M * Loki|muh hrhr
1105527625 M * Loki|muh should I rephrase my question? *g*
1105527635 M * Bertl yes, please ;)
1105527841 M * Loki|muh which version does not crash in normal use and has no important bugs?
1105527860 M * Loki|muh (thats what I consider as stable *g*)
1105527862 M * Bertl hmm, 1.29
1105527870 M * Loki|muh hu?
1105527880 M * Loki|muh 1.9.3 does not crash, either
1105527893 M * Bertl but it has important bugs ;)
1105527917 M * Loki|muh which?
1105527926 M * Bertl locking issue ...
1105527947 M * Bertl barrier deficiencies
1105527968 M * Bertl issues with the legacy calls and internal xid tagging
1105528153 M * Loki|muh what impact does the first 2 have?
1105528176 M * Bertl the locking issues might crash your kernel on SMP machines
1105528195 M * Bertl the barrier doesn't work for legacy vservers
1105528305 M * Loki|muh fine, I have no smp, no legacy things and no unification :)
1105528317 M * Loki|muh so let my rephrase my question once again
1105528323 M * Bertl so it works for you then ;)
1105528348 M * Loki|muh what version newer that 1.9.3 works as good as 1.9.3? Is there any?
1105528365 M * Bertl on your setup? sure, 1.9.3.17 should be fine ...
1105528393 M * Loki|muh thanks :)
1105528404 M * Bertl my pleasure!
1105528788 M * Loki|muh anything I should take care of when upgrading?
1105528838 M * Bertl hmm, no should be fine ...
1105529203 M * Bertl okay, back later ...
1105529210 N * Bertl Bertl_oO
1105530048 J * cabane cabane@m24s05.vlinux.de
1105530059 M * cabane hello 
1105530072 M * Loki|muh hi cabane 
1105530093 M * cabane I'
1105530102 M * cabane or more like more than one ;-)
1105530164 M * cabane I've got a debian host with kernel w/ vserver support and I've installed one vserver
1105530184 M * cabane apache on the vserver works as expected
1105530198 M * cabane I'm not able to ssh to the server, though
1105530211 M * cabane sounds familiar, I suppose
1105530255 M * Doener cabane: you have to make the host's sshd listen to only a limited set of ip addresses instead of 0.0.0.0
1105530261 M * cabane thanks
1105530288 M * cabane I tried a different approach - sshd for the host listens to port 222
1105530317 M * Doener possible as well, but that makes port 222 unusable in the vservers ;)
1105530324 M * cabane right
1105530338 M * cabane still I can't connect
1105530359 M * cabane first I could connect (to the host) now I can't connect at all :(
1105530369 M * Doener is the sshd in the vserver running? (it probably refused to start when you started the vserver)
1105530388 M * cabane vtop shows the sshd
1105530402 M * cabane and I stopped/started the vserver, too
1105530411 M * Doener try: chcontext --ctx 1 netstat -tnlp | grep sshd
1105530426 M * cabane ah! that'
1105530464 M * cabane I was about to ask, how it is possible to view/modify a vserver's file system
1105530499 M * Doener 2.4 or 2.6 kernel?
1105530503 M * cabane 2.4
1105530571 M * Doener well, the filesystem is just a (security enhanced) chroot. so it's pretty easy to view/change ;)
1105530612 M * cabane hoe do I do that? (bloody beginner's question)
1105530616 M * Doener i asked about the kernel version, because linux-vserver on 2.6 allows the usage of namespaces which require some additional work to modify a vserver's mount tree, but for 2.4 this isn't necessary
1105530630 M * cabane ok
1105530668 M * cabane is the context changed using the vserver's name?
1105530734 M * Doener cabane: are you working for vlinux.de ?
1105530770 M * cabane no :)
1105530791 M * cabane I'm from unilang.org
1105530799 M * cabane non-profit
1105530808 M * cabane www.unilang.org 
1105530842 M * cabane I'm in the office of my daytime job now - lunch break 
1105530855 M * Doener ok, just asking because i guess they should change their product descriptions... -,99 Euro/GB looks EXPENSIVE at first glance, because one (or at least I) does not really recognize the -, but simply reads 99 Euro/GB ;)
1105530870 M * cabane lol
1105530888 M * cabane I'
1105530943 M * nox expensive nevertheless 25 eur *pff*
1105530952 M * cabane actually he has a great produkt - the installation of the vserver-patched kernel on our new dedicated machine came for free
1105530971 M * cabane 5.80 EURO with 25GB
1105531005 M * cabane that's what we pay for our backup-vserver 
1105531030 M * cabane the real machine we get now costs 23EURO
1105531039 M * nox thats ok
1105531066 M * cabane we want vserver's for security reasons and because we want to have a production and a testing system
1105531111 M * nox i have a 1TB server for 30 but extra ip 5  eur  , godd that local-nat worx *g
1105531127 J * pusling_ ~pusling@195.215.29.124
1105531152 Q * pusling Read error: Connection reset by peer
1105531152 M * cabane wow
1105531239 N * pusling_ pusling
1105531250 M * cabane I need to leave now... I still have a lot of questions - maybe tonight!
1105531250 M * cabane bye
1105531255 Q * cabane Quit: using sirc version 2.211+ssfe
1105533551 Q * amit Quit: using sirc version 2.211+KSIRC/1.3.11
1105533678 M * TheSeer nice... yet another root exploit on linux ,>
1105534218 P * frz 
1105534697 M * TheSeer ensc: you awake? *g+
1105535311 M * TheSeer ensc: got it ;>  
1105535763 N * Bertl_oO Bertl
1105535783 M * Bertl TheSeer: hmm?
1105535809 M * TheSeer i was trying to get a fc3 vserver installed
1105535821 M * Bertl didn't work out?
1105535837 M * TheSeer the vserver-build script didn't find the repositories
1105535856 M * TheSeer because i forgot to update the apt stuff in the /etc/vservers folder
1105535878 M * TheSeer i was just wondering where else i might have to make changes.. but i got it working by now ;)
1105535904 M * Loki|muh TheSeer: is there a howto?
1105535917 M * Loki|muh if not, can you write one in the wiki? ;)
1105535935 M * TheSeer Loki|muh: hehe.. sure can.. i was planning on writing a howto for fc2/fc3 anyway..
1105535983 M * Loki|muh :)
1105535988 M * TheSeer Loki|muh: but telling util-vserver about fc3 is hardly more then copying all fc2 entries to fc3 and modifiy the now fc3 apt/sources.list to point to fc3 rather then fc2
1105536000 M * TheSeer you just have to find all of them ;)
1105536048 M * Loki|muh hehe
1105536138 M * TheSeer erm.. how do i use vrpm???
1105536155 M * TheSeer vrpm db -- -ivh MySQL-*.rpm -> no such file or directory
1105536160 M * TheSeer ?
1105537434 M * TheSeer hmm.. somehow vrpm must be broken.. :/
1105537564 J * tchan_ ~tchan@c-24-13-81-164.client.comcast.net
1105537588 Q * tchan Killed (NickServ command used by tchan_)
1105537602 N * tchan_ tchan
1105538113 M * brc BERTL
1105538118 M * brc ?
1105538118 M * brc ltns 
1105538118 M * brc heheh
1105538288 M * Bertl yep?
1105538455 M * brc Is thre a new version for 2.6.10 ?
1105538464 M * Bertl hmm, 1.9.3.17?
1105538482 M * brc i owe you some tests, sorry i have been really busy
1105538484 M * brc do you still need them ?
1105538496 M * brc frequency tests on a real computer to check smoothness :)
1105538516 M * Bertl ah, well, you can do them if you like, but they are not really important ...
1105538555 M * Bertl it's more to get some opinions on the usability ...
1105538576 M * Bertl i.e. if this _is_ something to pursue in the future ...
1105538650 M * brc ok
1105538668 M * brc 2.6.9-vs1.9.3 has been really stable for me.
1105538674 M * brc almost 60 days uptime
1105538682 M * Bertl great! SMP or UP?
1105538687 M * brc UP
1105538694 M * brc the only problem is with context "locking"
1105538702 M * brc i stop a vserver, when i try to stop, it says "context allready exists"
1105538712 M * brc that happens on all of my 3 servers
1105538715 M * pulsar I have SMP uptime about 30 days on 2.6.9-vs1.9.3
1105538716 M * brc so i change the context and start
1105538730 M * pulsar 12 VPS's
1105538742 M * Bertl brc: i.c. did you try the mount -o remount,rw /proc trick yet?
1105538753 M * Bertl pulsar: hmm sounds good!
1105538803 M * brc now yet
1105538804 M * brc gonna ty it
1105538940 M * pulsar Bertl, UP -> Linux sls-eb7p4 2.6.9-vs1.9.3-rc5 #1 Mon Nov 1 21:07:14 EET 2004 i686 i686 i386 GNU/Linux
1105538941 M * pulsar [root@sls-eb7p4 root]# uptime
1105538941 M * pulsar  16:03:56 up 71 days, 18:49,  1 user,  load average: 0.19, 0.18, 0.16
1105538962 M * pulsar stable for me
1105539238 M * Bertl excellent!
1105539281 M * Bertl I hope we can increase the stability further ...
1105539804 M * TheSeer .6.9-vs1.9.3-rc5 #1 SMP Sun Oct 31 17:32:57 CET 2004
1105539804 M * TheSeer [root@core root]# uptime
1105539804 M * TheSeer  15:35:01 up 49 days, 
1105539807 M * TheSeer looks good here too ;)
1105539829 M * TheSeer 2.6.9 that is of course ;>  sorry for the bad paste..
1105539896 M * Bertl yeah, 2.6.9 had the advantage to follow the 'strange' 2.6.8.1 release after a longer cleanup period ...
1105539919 M * Bertl I expect 2.6.11 to be of similar stability ...
1105539948 M * TheSeer they have quite a few bad bugs to fix for that release...
1105540211 M * TheSeer gna.. sometimes i don't get dependencies..
1105540227 M * TheSeer why the f*** do i need to install perl-DBD-stuff to install php-mysql?
1105540228 M * TheSeer *grmpf*
1105540240 A * TheSeer kicks the fedora rpm package builder crew...
1105540256 M * Bertl well, recent mandrake have deps like this:
1105540300 M * Bertl sgml-common -> latex -> gtk -> x11
1105540309 M * brc Bertl: Is 2.6.10 much more stable then 1.9.3?Is it worth upgrading?
1105540318 M * TheSeer brc: ??
1105540326 M * Bertl second that! ??
1105540346 M * TheSeer Bertl: so you need x11 to use sgml.. kewl :)
1105540407 M * Bertl I really wonder when the first rpm will depend on the kitchen-sink ;)
1105540433 M * TheSeer hehe
1105540585 M * brc oops
1105540589 M * brc Bertl: Is 2.6.10 much more stable then 2.6.9?Is it worth upgrading?
1105540615 M * Bertl nope, 2.6.10 seems to be _less_ stable than 2.6.9
1105540740 M * brc weird. why ?
1105540750 M * brc i though they were fixing the bugs, but it seems they are creating new ones!!!
1105540765 M * Bertl well, there are 3 components ...
1105540780 M * Bertl 1) bugfixes which just fix bugs ...
1105540796 M * Bertl 2) bugfixes or improvements which break other stuff
1105540807 M * Bertl 3) new features which are introduced
1105540846 M * Bertl depending on the mix, you increase or decrease stability ;)
1105540901 M * brc hehe sux
1105540911 M * pulsar 2.6.9 & 2.6.10 have root exploitable holes , but exploits does not work on vps. Anyway ... waiting for 2.6.11
1105540931 M * Bertl hmm, you tried that?
1105540932 M * brc didn't know about that. have't read bugtraq for a while.
1105540944 M * pulsar Bertl, yes
1105540972 M * Bertl didn't read a mail on the ml about that ... would have been interresting for folks out there ...
1105541203 J * pflanze ~chris@unk-110.ethz.ch
1105541212 M * pflanze Hello
1105541213 M * Bertl welcome pflanze!
1105541244 P * pflanze 
1105541245 J * pflanze ~chris@unk-110.ethz.ch
1105541249 M * pflanze Hello
1105541254 M * Bertl hello again!
1105541256 M * pflanze :)
1105541266 M * pflanze I just wanted to test whether you had running a bot.
1105541283 M * pflanze At least now I know that it would have be an intelligent bot.
1105541305 M * pflanze You're just too fast responding :)
1105541318 M * Bertl so tell me about your problems, chris! *G*
1105541346 M * pflanze No problems (this time).
1105541370 M * Bertl hey, this was my eliza mode ;)
1105541375 M * pflanze Just a question: how do I determine the load which each vserver is causing.
1105541479 M * mugwump start with vserver-stat
1105541536 M * Bertl hmm, what about  grep loadavg /proc/virtual/<xid>/cvirt ?
1105541541 M * pflanze Here is the output from vserver-stat http://elvis-jaeger.mine.nu/load
1105541589 M * pflanze userTIME and sysTIME is there, but I even don't know if those are just the sum of all times of the currently-running processes.
1105541598 A * pflanze going to look at the cvirt files
1105541608 M * Bertl scrat LOL!
1105541660 M * pflanze that's the one of one of my friends, www.metzner.org
1105541756 M * pflanze I don't have <xid>/cvirt files.
1105541763 M * pflanze I guess it's time to upgrade to 2.6..
1105541767 M * Bertl ;)
1105541973 M * Bertl hey mugwump, btw!
1105542019 M * Bertl mugwump: we declared this the year of linux-vserver, are you going to help? ;)
1105542484 M * pflanze The year of linux-vserver? :)
1105542505 M * Bertl yes, world domination, you know?
1105542514 M * pflanze heh:)
1105542588 M * pflanze Why not? 
1105542673 J * Brucey cinacio@snape.aspic.com
1105542683 M * Brucey hi guys
1105542689 M * Brucey (and girls)
1105542717 M * Bertl hey Brucey!
1105542723 M * pflanze Hi Brucey. You should say 'Hello' to trigger Bertl's eliza mode.
1105542749 M * Loki|muh hrhr
1105542797 M * Snow-Man vservers and grub get along fine, don't they?
1105542814 M * Bertl at least on my hosts ;)
1105542823 M * Snow-Man Have any amd64 machines?
1105542836 M * Brucey pflanze: I'd rather only trigger his "awake" mode :)
1105542876 Q * sannes Read error: Connection reset by peer
1105542887 M * Bertl Snow-Man: nope, do you need a postal address? ;)
1105542894 M * Loki|muh Snow-Man: amd64 works fine with vserver except latest alpha-tools
1105542915 M * Snow-Man Loki|muh: Just having some problems w/ grub.
1105542926 M * Snow-Man On amd64, on a machine that happens to have a vserver kernel.
1105542935 M * Snow-Man I'm not sure that it's the vservers tho, just a thought.
1105542943 M * Snow-Man Loki|muh: Latest being 0.30.196?
1105542973 M * Brucey I don't see the difference a vserver enabled kernel would do to grub
1105542974 M * ndim Are you running grub on the host or the guest? The host should not make any problems, and the guest should not have access anyway.
1105542993 M * Snow-Man ndim: On the host, there aren't actually any guests yet.
1105542999 M * Loki|muh Snow-Man: yeah, 0.30.196 works not, but 0.30.195 works fine
1105543003 M * Snow-Man Brucey: <Shrug> It was just a thought, I suspect it's probably a 2.6.10 issue. :)
1105543009 M * Snow-Man Loki|muh: Well, that sucks.
1105543019 M * ndim What is the problem with 0.30.196?
1105543021 M * Snow-Man Loki|muh: Is there a patch or something to fix it
1105543021 M * Snow-Man ?
1105543031 M * Snow-Man Yeah, and what's the problem, exactly? :)
1105543040 M * Brucey Snow-Man: just out of curiosity, what's the symptom of your problem?
1105543057 M * Snow-Man Brucey: grub segfaulting
1105543070 M * Loki|muh Snow-Man: the problem is that there is no implementation of nice in the dietlibc, imo
1105543087 M * Loki|muh Snow-Man: and if you compile dietlibc by yourself you need a patch for amd64, too
1105543088 M * Snow-Man Loki|muh: What if you're not using dietlibc?
1105543095 M * ndim Loki|muh: Using legacy config?
1105543095 M * Loki|muh i dunno
1105543112 M * Loki|muh ndim: no, nothing legacy here
1105543130 M * Snow-Man What's the reason to use dietlibc anyway?
1105543146 M * ndim resolver problems in chroot
1105543157 M * ndim (with glibc)
1105543173 M * Snow-Man I don't recall seeing any...?
1105543200 M * Snow-Man (resolver problems that is, in a vserver w/ Debian/sid)
1105543240 M * Snow-Man Should I be seeing some? :)
1105543389 M * Snow-Man hello?
1105543390 M * Seraph Loki|muh: why patch for amd64?
1105543403 M * Seraph Loki|muh: i have compiled util-vserver with dietlibc on amd64.. no problems..
1105543405 M * Bertl hello Snow-Man! ;)
1105543414 M * Snow-Man Bertl: heya
1105543416 M * Seraph Loki|muh: though i haven't tried running it ..
1105543456 M * ndim Snow-Man: No idea what exactly these problems are and where, but I suppose there is a reason behind it.
1105543491 M * Loki|muh http://doener.homeip.net/doener/dietlibc-amd64.diff
1105543496 M * Loki|muh this is the amd64 patch
1105543504 M * Loki|muh for the latest dietlibc release
1105543512 M * Brucey Snow-Man: grub segfaults when it tries to load the kernel?
1105543529 M * Brucey and it works with other kernels?
1105543599 M * Bertl grub segfaults?!
1105543610 A * Bertl is reading up the log ...
1105543638 M * Seraph Loki|muh: that's needed for compilation or for function?
1105543680 M * Bertl Seraph: the dietlibc seems to be broken on amd64 regarding size of some (especially env) pointers
1105543701 M * Seraph Bertl: then the debian folks have either fixed that or it does only happen on execution
1105543704 M * Seraph Bertl: but compilation was fine
1105543724 M * Bertl so if you get segfaults on some tools (when executing), then this is very likely to be the cause ...
1105543731 M * Seraph hmm, ok
1105543882 M * Snow-Man Brucey: No, it segfaults when it tries to load the stage2 bootloader on to the disk.
1105543894 M * Snow-Man ndim: Hrmpf, I don't like that.
1105543927 M * Brucey Snow-Man: then the kernel has nothing to do with it :)
1105543933 M * Snow-Man Bertl: Yes, grub is segfaulting.
1105543939 M * Snow-Man Brucey: uhm, it works fine under 2.6.8
1105543949 M * Snow-Man Brucey: The only thing that changes is the kernel.
1105543971 M * Bertl Snow-Man: hmm, didn't know that grub _can_ segfault ...
1105543982 M * Snow-Man Anything can segfault. :)
1105544011 M * Brucey weird... because when grub loads the stage2 bootloader it doesn't do anything with the kernel yet
1105544011 M * Snow-Man Bertl: Under 2.6.8 it can install the stage2 bootloader fine, under 2.6.10 it segfaults. :)
1105544017 M * Snow-Man Brucey: nono.
1105544021 M * Snow-Man Brucey: You're misunderstanding.
1105544028 M * Snow-Man After the system has BOOTED COMPLETELY
1105544030 M * Bertl ah, you are talking about the installation tool
1105544031 M * Snow-Man and you run grub
1105544033 M * Brucey oh... you're running the grub binary to install it?
1105544035 M * Snow-Man and do 'setup ....'
1105544039 M * Brucey oh okay
1105544040 M * Snow-Man It segfaults
1105544059 M * Bertl most of us probably thought that stage 1-2 segfaulted on you ;)
1105544061 M * Snow-Man Bertl: Yeah, when trying to install the stage2 bootloader.
1105544082 M * Snow-Man Bertl: No..  *That* would have been interesting.
1105544083 M * Bertl well, then get gdb installed and look _where_ it happens ...
1105544093 M * Snow-Man I did an strace on it
1105544104 M * Snow-Man I'm not too sure about trying to gdb it..
1105544116 M * Snow-Man This is on a Debian/pure64 box, and the grub binary is 32bit, and gdb is 64bit. :)
1105544125 M * Bertl you can 'gdb' the coredump ;)
1105544166 M * Snow-Man #81 0x08073d7a in ?? ()
1105544167 M * Snow-Man Previous frame inner to this frame (corrupt stack?)
1105544168 M * Snow-Man hate that.
1105544184 M * Snow-Man  Running "embed /boot/grub/e2fs_stage1_5 (hd0)"...  16 sectors are embedded.
1105544184 M * Snow-Man succeeded
1105544187 M * Snow-Man Program received signal SIGSEGV, Segmentation fault.
1105544212 M * Bertl okay, what kernel?
1105544223 M * Snow-Man 2.6.10, that's what I've been telling you. :)
1105544226 M * Snow-Man Works fine under 2.6.8
1105544231 M * Bertl 2.6.10 vanilla?
1105544252 M * Bertl 2.6.10 debian special?
1105544252 M * Snow-Man Nope, 2.6.10 w/ Debian patches, vserver and ipsec patches.
1105544271 M * Bertl k, then first try 2.6.10 debian
1105544283 M * Bertl then 2.6.10 vanilla + vserver ...
1105544299 M * Bertl if you get a segfault there, we have an issue ;)
1105544314 M * Snow-Man Well, that's what I was asking, if vserver could affect it..  I'd tend to think 'no', but thought I'd ask/
1105544326 M * Bertl IMHO no, not on the host
1105544332 M * Snow-Man I'm already bugging people about it on the debian-amd64 list.
1105544355 M * Bertl I would expect either 2.6.10 vanilla or debian patches ...
1105544373 M * Snow-Man I suspect 2.6.10, I think I saw something on lkml about this previously.
1105544397 M * Snow-Man I don't recall anything in the Debian patches that would be likely to affect this (really, the Debian patches aren't very big...)
1105544427 M * Bertl k, don't really know them ...
1105544440 M * Snow-Man Guess no one else has seen this yet?
1105544504 M * Seraph Snow-Man: well, running x86_64 with a 2.6.10 vanilla+vserver into an i386 system is so far without problems
1105544535 M * Seraph Snow-Man: yet i'm not sure vserver works as expected with the 32bit userland under 64bit linux-kernel
1105544594 M * Snow-Man Well, this is a 64bit userland and 64bit kernel. :)
1105545359 Q * virtuoso Remote host closed the connection
1105546069 J * virtuoso ~s0t0na@tranq.dorms.spbu.ru
1105546130 M * Bertl wb virtuoso!
1105546328 J * shuri ~shuri@dsl.speedline209.226.electronicbox.net
1105546356 M * Bertl hello shuri!
1105546364 M * shuri hola Bertl 
1105548676 Q * pulsar Remote host closed the connection
1105548944 J * el ~el@229-154-088-212.ip-addr.teresto.net
1105548946 M * el hi
1105548951 M * Bertl welcome el!
1105548991 M * el Hi Bertl. (I hope you are not a bot)
1105549007 M * Bertl hmm, well, what if?
1105549083 M * el I have a strange problem with vserver 1.29. If I entered a vserver i see all network interfaces and the vserver uses the eth0 interfase intead of eth0:vserver name. This is since the upgrade to 1.29
1105549115 M * el Bertl: bot.. (because of your fast answer)
1105549142 M * Bertl wouldn't a smart bot use some _long_ delay then ;)
1105549159 M * el hmm, you are right ;)
1105549178 M * Bertl and a smarter one avoid just that .. to make sure that you do not think it's a bot ;)
1105549216 M * Bertl btw, did you update anything else except the kernel/vserver patch?
1105549223 M * el perhaps, you are a very simple bot ;)
1105549242 M * Bertl hmm, right, never considered that ... 
1105549282 M * el ok, back to my problem. i have also change the ulimit from the config file to ULIMIT="-H -n 1024"
1105549355 Q * BWare Quit: using sirc version 2.211+KSIRC/1.3.10
1105549385 M * el that is everything I have change. i think i'm not the fist person how upgraded to 1.29 and came into that trouble!?
1105549393 M * Bertl that's all, not maybe the tools too?
1105549420 M * el oh, i have to check, I use debian sarge.
1105549502 M * el hmm, the vserver tools where upgraded from util-vserver 0.30-8 to 0.30-10
1105549555 M * Bertl now as I have no idea what changed between those releases ... I'd say, let's try to revert that change and check again ...
1105549633 M * ntrs Is there any way to determine which vserver is using the most memory/swap. Something sort of like vmstat with real-time output but vserver specific.
1105549660 M * ntrs Or at least which vservers is trashing and why?
1105549710 M * Bertl not really, but if you find a way to tell which 'process' is trashing, let me know, and we can determine the context ...
1105549745 M * Bertl why is relatively easy, because system memory is to small and page in/out takes too long ...
1105549784 M * ntrs I was hoping you could tell me that there was a way to tell which process is trashing. :)
1105549839 M * Bertl hmm ...
1105549894 M * ntrs ok, what do you think about upgrading the memory from 2 to 4 GB?
1105549929 M * Bertl basically a bad idea when you are on x86 IMHO
1105549946 M * ntrs So, there is pretty much no solution, right?
1105549959 M * Loki|muh Bertl: why is this a bad idea?
1105549964 M * Bertl ntrs: depends on the issue ...
1105549965 M * ntrs except using different architecture?
1105549988 M * Bertl Loki|muh: because you then require highmem, which adds overhead, etc ...
1105549994 M * ntrs well, obviously the server is swapping too much, that is the issue
1105550009 J * sannes ~ace@home.skarby.no
1105550009 M * Loki|muh ah, thx
1105550018 M * Bertl welcome sannes!
1105550018 M * el Bertl: no,i have the same problem with the old util-vserver tools with the old util-vserver tools
1105550019 M * ntrs it will probably add less overhead than the server being in IOWAIT state about 95% of the time
1105550086 M * Bertl el: k, then let's check a few things ... first the testme.sh (http://vserver.13thfloor.at/Stuff/SCRIPT/testme.sh)
1105550119 M * Bertl ntrs: IOWAIT is a good indication for a misbehaving process ...
1105550128 M * Bertl (or bad IO scheduling)
1105550148 M * Bertl (extremely high and constant IOWAIT that is)
1105550217 M * el Bertl: chcontext is workting, chbind is working, everything succeeded/
1105550232 M * Bertl okay, that is fine ...
1105550272 M * Bertl now could you upload your vserver config (replace sensitive info) somewhere (i.e. pastebin.com)?
1105550363 M * el ah, Bertl on my old system i have a proc and dev directory off die vservers dir. know in the verservers dir is only die vservers-root.
1105550453 M * el Bertl: http://nopaste.php-q.net/107813
1105550475 M * Bertl tx
1105550477 Q * grecea Read error: Connection reset by peer
1105550550 M * Bertl el: any reason for the various flags and the cap?
1105550569 M * Bertl (just curious)
1105550580 M * el Bertl: hmm, it's debian default
1105550620 M * Bertl well, I hope that will change soon ... k
1105550735 M * Bertl and when you start the vserver, what messages do you get?
1105550922 J * grecea ~grecea@h-195-22-237-74.mdl.net
1105550935 M * Bertl welcome grecea!
1105550939 M * el Bertl:  http://paste.phpfi.com/45591
1105550946 M * Bertl tx
1105550953 M * el hi grecea
1105550980 M * el /var/run is no symlink.
1105550996 M * el oh, without the fist /
1105551021 M * Bertl hmm, how did that happen?
1105551127 M * el ?
1105551137 M * Bertl looks like something in the vserver setup changed and it probably breakes the startup config ... which explains the missing alias ...
1105551139 M * el I don't understand what you mean
1105551164 M * el k
1105551174 M * Bertl did you get the Symlink messages before?
1105551196 M * el no
1105551216 M * Bertl hmm, do you still have the old kernel available?
1105551233 M * Bertl (the one you used fbefore the upgrade)
1105551244 M * el yes. the config. I have to compile the old kernel
1105551274 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il
1105551275 M * Bertl k, would be interesting to see if it is kernel related at all ... personally I doubt it ...
1105551279 M * Bertl welcome flock!
1105551288 M * el hi flock 
1105551308 M * flock 'lo guys. How's stuff?
1105551345 M * Bertl lying around everywhere :/
1105551445 M * ndim piling towards the heavens :)
1105551695 Q * Brucey Quit: "War doesn't determine who's right, war determines who's left."
1105552128 Q * rs Quit: leaving
1105553335 Q * Loki|muh Read error: Connection reset by peer
1105553342 J * Loki|muh loki@satanix.de
1105553760 M * el ok, I have to go home. thx Bertl I will report tomorrow
1105553774 M * Bertl ok, cya!
1105553779 Q * el Quit: leaving
1105554021 Q * Loki|muh Ping timeout: 480 seconds
1105554132 J * nayco ~nayco@lns-vlq-47-nan-82-252-230-84.adsl.proxad.net
1105554139 M * nayco 'llo !
1105554249 J * Loki|muh loki@satanix.de
1105554266 M * Bertl hey nayco!
1105554470 J * mhepp ~mhepp@r72s22p13.home.nbox.cz
1105554561 M * nayco Bertl: I noticed something strange: I managed to install a template vserver with "urpmi --root..." (Run from the host) and I get a ~100Mb install. Good. Then, I used util-vserver-0.30 to lauch it, it runs badly because of the lack of "vprocunhide"... So I tried with util-vservers-0.30-196, and it starts better, but with no IP (The interface alias isn't Created/started...) Do you know why ?
1105554615 M * Bertl first, did you config an alias at all?
1105554636 M * Bertl (the default is to _not_ use aliases for the ips)
1105554654 M * nayco Errrr, no, but it is not needed with 0.30, as I saw (Or was I lucky ?)
1105554676 M * Bertl checking with ip addr ls  (inside the vserver, or on the host) will help you there
1105554711 M * nayco With 0.30, after starting the vserver, I get automatically an alias "etho:test" created...
1105554737 M * Bertl yes, that is expected, but the alpha tools use a different kind of setup (as default)
1105554765 M * Bertl you can choose to create an alias, but you have to do so explicitely
1105554775 M * nayco Well, inside the vserver, I used "ifconfig": Nothing. With "ifconfig eth0", I get 2-3 lines with MAC, IPV6 adress...
1105554804 M * nayco And if I do not create an alias ? It runs on the main interface ?
1105554844 M * Bertl it always 'runs' on the 'main' interface ... just the way the additional ips are declared is different
1105554867 M * Bertl (ifconfig is an old tool, it doesn't see everything, use ip for that)
1105554956 M * nayco so you mean that even with aliases, I can connect my vserver on a given port ? Are there conflicts with the main host if so ? (Well, the answer seems obvious, but I'm confused, here ;-))
1105554973 M * nayco ok for ifconfig/ip.
1105555033 M * Bertl yep, (without ngnet) an app binding to 0.0.0.0 on the host for example will _always_ clash with any vserver (trying to bind the same port)
1105555039 M * nayco Anyway, the stable tools work good for me, is there a way to get the same behaviour that with vprocunhide with this realease ? "setattr" is a bit hard for me...
1105555091 M * Bertl you basically have two options:
1105555097 M * nayco So, it is best (Easier/safer ? ) to use aliases, so ?
1105555104 M * nayco yes ?
1105555113 M * Bertl a) you can use the vproc tool and do what vprocunhide does by hand
1105555149 M * Bertl b) disable the vproc security (which isn't worse than 2.4/1.2x without a proper script) but it _is_ a security issue if you have untrusted vserver root
1105555186 M * Bertl ad aliases vs ips: neither method is safer or easier ...
1105555204 M * Bertl (and of course, you can use aliases with alpha too, as I said)
1105555293 M * Bertl okay, moving out now, back in a few hours ...
1105555303 N * Bertl Bertl_oO
1105555387 Q * Loki|muh Read error: Connection reset by peer
1105555494 J * Loki|muh loki@satanix.de
1105555619 M * nayco thanks !
1105555645 Q * serving_ Ping timeout: 480 seconds
1105555698 Q * shuri Remote host closed the connection
1105555869 J * frz ~frz@jaim.at
1105555917 P * frz 
1105556707 Q * Loki|muh Read error: Connection reset by peer
1105556735 J * Loki|muh loki@satanix.de
1105557011 N * chris|sleep chrish01
1105558577 M * chrish01 hey guys
1105558770 Q * pusling Ping timeout: 480 seconds
1105558997 J * pusling ~pusling@195.215.29.124
1105559056 Q * chrish01 Quit: off to work
1105559777 Q * we2by Read error: Connection reset by peer
1105559777 Q * Loki|muh Read error: Connection reset by peer
1105559824 J * Loki|muh loki@satanix.de
1105560590 J * we2by ~we2by@dc5146d009.adsl.wanadoo.nl
1105560865 J * chrish01 ~chrish01@69.90.131.10
1105560888 M * lilo semi
1105560893 M * lilo oops ;)
1105560914 Q * lilo Quit: hmmm, udev
1105561029 Q * sannes Read error: Connection reset by peer
1105561267 Q * Loki|muh Read error: Connection reset by peer
1105561320 J * Loki|muh loki@satanix.de
1105561391 Q * pusling Read error: Connection reset by peer
1105561391 J * pusling_ ~pusling@195.215.29.124
1105561402 J * lilo ~lilo@lilo.usercloak.oftc.net
1105561702 Q * nox jupiter.oftc.net neutron.oftc.net
1105561845 J * nox ~nox@c207113.adsl.hansenet.de
1105562050 Q * nox jupiter.oftc.net neutron.oftc.net
1105562077 J * nox ~nox@c207113.adsl.hansenet.de
1105562322 M * chrish01 Bertl_oO, when you get back in...let me know if you have done any work on NGNet with vservers
1105562510 J * serving ~serving@213.186.172.180
1105562547 Q * Loki|muh Read error: Connection reset by peer
1105562550 Q * mhepp Quit: mhepp caught signal: Autobus error
1105562600 J * Loki|muh loki@satanix.de
1105562614 Q * nox jupiter.oftc.net neutron.oftc.net
1105563237 J * nox ~nox@c207113.adsl.hansenet.de
1105563787 Q * Loki|muh Read error: Connection reset by peer
1105563813 J * Loki|muh loki@satanix.de
1105564285 J * sannes ~ace@home.skarby.no
1105564688 M * sannes :)
1105565038 Q * Loki|muh Read error: Connection reset by peer
1105565042 J * Loki|muh loki@satanix.de
1105567146 Q * Loki|muh Read error: Connection reset by peer
1105567200 J * Loki|muh loki@satanix.de
1105567637 J * thresh ~popa3d@cryo.net.ru
1105567686 M * thresh hi. i've got a problem with openpty issue with sshd under vserver. vserver version is 1.28
1105567702 M * thresh when i login over ssh to my vserver it says in logs:
1105567705 M * thresh Jan 11 18:52:55 cryo sshd[14457]: error: openpty: No such file or directory
1105567705 M * thresh Jan 11 18:52:55 cryo sshd[14464]: error: session_pty_req: session 0 alloc failed
1105567711 M * thresh what could be the problem?
1105568367 M * nayco ohhh, I read it from the faq this afternoon.... sh***, i can't remember what is the problem :(
1105568385 M * nayco ...something to do with /dev/pty...
1105568396 M * thresh what's the url of faq? :)
1105568396 Q * Loki|muh Read error: Connection reset by peer
1105568416 M * nayco ....wait ;-)
1105568474 M * Doener thresh: is devpts mounted? are unix98 ptys activated in your kernel configuration?
1105568489 M * thresh Doener: devpts is mounted in host system
1105568500 M * thresh in /proc/mounts i see:
1105568514 J * Loki|muh loki@satanix.de
1105568514 M * thresh thresh@cryo ~ $ cat /proc/mounts| grep pts
1105568514 M * thresh devpts /dev/pts devpts rw 0 0
1105568514 M * thresh none /var/lib/vservers/cryo.net.ru/dev/pts devpts rw 0 0
1105568524 M * Doener ok, so it's mounted in the vserver
1105568547 M * Doener what does your vserver's dev directory look like?
1105568599 M * thresh [root@vserver:cryo.net.ru dev]ls /dev/
1105568599 M * thresh log  null  pts  tty12  urandom  zero
1105568653 M * nayco needs a makedev ?
1105568670 M * Doener ok, do: cd /dev; cp -a full null ptmx random tty urandom zero /path/to/vserver/dev
1105568681 M * Doener nayco: vservers are not allowed to mknod ;)
1105568683 M * thresh ooh. ok.
1105568704 M * nayco Doener: Well, in fact, that's better like that ;)
1105568734 M * thresh seems like i hadn't read the docs well :)
1105568769 M * Doener btw the cp is copying some file that are already there just because i've been too lazy to check which ones you already have ;)
1105568769 M * nayco Well, I cannot find this Faq with a question about pty... Sorry.
1105568795 M * thresh Doener: i know, i'd created null, zero, urandom, random before :)
1105568798 Q * Loki|muh Quit: 
1105568818 M * Doener let me know if it works, when you're ready
1105568830 M * thresh it works :)
1105568834 M * Doener great!
1105568843 M * Doener guess it was the missing ptmx
1105568855 M * thresh i'll check
1105568866 M * Doener i'm even pretty sure ;)
1105568889 M * thresh yes, it is :)
1105568900 M * thresh thank you :)
1105568905 M * Doener you're welcome
1105568962 M * nayco By the way, Is it a good idea to have devfs mounted in the vservers dev dir, or are the default entries of the distro's basesystem sufficient ?
1105569224 J * boklm boklm@woper.mars-attacks.org
1105569226 M * boklm hello
1105569249 Q * chrish01 Remote host closed the connection
1105569258 J * Loki|muh loki@satanix.de
1105569296 M * boklm http://www.isec.pl/vulnerabilities/isec-0021-uselib.txt
1105569319 M * boklm does anyone know what users can do if users exploit this hole from a vserver ?
1105569336 M * boklm I wonder if it's easy to get out of the vserver
1105569359 M * Loki|muh bertl stated something that uml is vulnerable but vserver not
1105569369 M * Loki|muh dunno if it is that hole or another
1105569378 M * boklm ok
1105569945 N * pusling_ pusling
1105570133 M * Doener nayco: in a vserver you should have only a very limited set of files in /dev
1105570186 M * Doener f.e. having sda1 in a vserver would give the vserver's root the possibility to access that partition and f.e. destroy data...
1105570277 Q * Loki|muh Read error: Connection reset by peer
1105570331 J * Loki|muh loki@satanix.de
1105570432 Q * Loki|muh Quit: 
1105571716 M * nayco Doener: Ok. In fact, you mean the entries created by "vserver ... build -m skeleton" are sufficient ?
1105571733 M * Doener exactly
1105571755 M * nayco Sufficient for most daemons or servers ?
1105571791 Q * jsambrook Ping timeout: 480 seconds
1105571859 M * Doener sufficient for about everything a vserver should do... you basically just lack immediate hardware access (which is bad anyways and not needed) + kmem/kmsg
1105571901 M * nayco ok, so httpd, mysql, ftpd, sshd, postfix for exemple will work fine ?
1105571907 J * chrish01 ~chrish01@69.90.131.10
1105571929 M * Doener of course! ;)
1105572002 M * nayco that is the way I plan to use vservers... Oh, I forgot, maybe the most important: I'd like a _safe_ on-demand environnement to compile packages for another distro...
1105572187 M * nayco I mean, I recently had to backport (Well, actually just tried to compile it praying that the dependancies were ok ;-)) a package from mdk 10.1 to 9.2, and had to do it on the target host... Prod mailserver :( It worked perfectly for 10.0-> 9.2, but failed from 10.1, so I had to give up an leave an old version running.
1105572225 M * nayco Didn't want to play with murphy in trying harder ;-)
1105572299 M * daniel_hozac fedora.us uses (used?) vservers to build packages ;)
1105572392 M * nayco That's axactly I want to do, one vserver for each distro version, to get last software... No to play geek, but some soft nned it (Fe: ClamAV, complaining during DB updates that its version is too old :( )
1105572591 J * prae ~prae@sherpadown.net
1105572594 M * prae hi all :)
1105572608 M * Doener welcome prae 
1105572614 M * prae thx :)
1105572784 M * nayco 'llo !
1105572826 M * chrish01 does anyone help bertl on ngnet?
1105572842 M * nayco what is ngnet ?
1105572867 M * chrish01 the next generation networking stuff...really cool network segmentation between context stuff
1105572909 M * nayco so the vservers cannot use the same networking device (Do they can, in fact ?)
1105572948 M * chrish01 you can do some really cool stuff...unfortunately only with contexts right now (im trying to pay him to do vservers too)
1105572962 M * chrish01 but you can map real devices to context devices and stuff
1105572970 M * chrish01 like segmented routing tables, iptables, etc...
1105573025 M * chrish01 so i can take say... eth0 and map a context device from it as en0 or something
1105573035 M * nayco Well, I think my poor vserver newbie knowledge prevent me to understand all ;-)
1105573054 M * nayco prevents
1105573132 M * chrish01 its a lot like how you would map devices in UML using taps
1105573142 M * chrish01 but sooo much faster since you dont have uml overhead
1105573263 M * prae what is advantages of vserver comparated to UML ?
1105573307 M * chrish01 overhead
1105573313 M * nayco so, now, with the current releases, I couldn't really see what happens with networking (Not done enough tests/script reading). Well, if my host has one eth (0), I use aliases for my vservers, so they are independant. But Bertl_oO told me earlier that it's not mandatory... That's why I don't really understand: The vservers can bind directly to eth0 too ?
1105573323 Q * thresh Quit: when i grow up there will be a day when everybody have to do it MY WAY
1105573338 M * chrish01 nayco, ngnet is completely different from anything you know about vserver now
1105573345 M * chrish01 anyway...afk brb
1105573421 M * nayco well, that's the problem... I still don't know much of the vserrvers ;-)
1105573780 M * chrish01 k back
1105573799 M * chrish01 you know how virtual routers work (like from cisco, alcatel)?
1105573806 M * nayco nope.
1105573840 M * nayco I just worked on a "real" router once (alcatel)
1105573843 M * chrish01 ah, you can read up if you want...im doing something similar now with a bunch of UML's ontop of a linux host...id like to do that with ngnet in the future cause its tons faster
1105573905 M * nayco So, what is a virtual router ?
1105573998 M * chrish01 one sec...
1105574044 M * nayco k :)
1105574076 M * chrish01 ok, you understand how VLANS work?
1105574081 M * nayco yes !
1105574082 M * chrish01 ok
1105574088 M * chrish01 its the same sort of deal...
1105574093 M * nayco mmm ?
1105574097 M * chrish01 in my one router...i can have say 1024 virtual routers
1105574106 M * chrish01 each has its own routing tables, firewalls, etc...
1105574121 M * chrish01 then i can pipe circuits (ethernet, t1, atm) into those virtual routers
1105574148 M * chrish01 that way, i can have this one router route packets for saw 100 customers :)
1105574152 M * chrish01 err 1000
1105574162 M * nayco ....
1105574202 M * nayco well... Let me understand.... Can this be done on two interfaces ?
1105574215 M * chrish01 SBS (RedBack) does that, but those routers are like $200,000+
1105574230 M * nayco :-$
1105574239 M * chrish01 well, lets say i have a VLAN trunk coming into this vserver host
1105574248 M * nayco yes
1105574257 M * chrish01 and each customer has say 10 - 100 circuits, each circuit on its own vlan
1105574274 M * nayco what do you call circuits ?
1105574298 M * chrish01 each circuit could be a t1, atm (dsl), or ethernet coming in
1105574301 M * chrish01 (we are an isp btw)
1105574324 M * nayco you mean _real_ rj45 cable for example ?
1105574340 M * chrish01 could be, but you are thinking too much right now
1105574346 M * nayco :D
1105574348 M * chrish01 think protocol, not hardware
1105574359 M * nayco i'll try