1136937724 M * daniel_hozac morning
1136937837 M * daniel_hozac have you thought about a vps-initscripts package?
1136937923 M * daniel_hozac (or do the utils solve that problem...)
1136938030 M * Cru the vps-fakepackages fake everything that is required by the original initscripts
1136938041 M * Cru required but useless
1136938080 M * daniel_hozac i meant more along the lines of removing /etc/rc.d/rc[06].d/S01{halt,reboot} and running chkconfig network off.
1136938098 M * daniel_hozac and maybe adding a real killall script.
1136938108 M * daniel_hozac but it looks like the utils might actually do that.
1136938164 A * daniel_hozac has always just used yum/rpm directly to build chroots.
1136938248 M * Cru hmm
1136938309 M * Cru with vs2.0 I experienced no init problems until now...
1136938328 M * Cru vserver-build seems to do this just fine for me
1136938432 M * daniel_hozac indeed.
1136938438 M * daniel_hozac guess i'll have to start using that instead.
1136939272 M * lchvdlch Bertl_oO: I hope the "works better" isn't ide/ata related :)
1136939381 Q * mnmr Ping timeout: 480 seconds
1136940580 Q * jeeves Quit: Leaving
1136941185 M * Cru bbl
1136941189 Q * Cru Quit: use Unices; $live->free() || die;
1136943255 Q * michal_ Ping timeout: 480 seconds
1136943394 J * michal_ ~michal@mprivacy-update.de
1136944466 Q * zobel Ping timeout: 480 seconds
1136944667 Q * eyck Remote host closed the connection
1136944855 J * zobel zobel@zobel.irc.ftbfs.de
1136945548 P * undefined 
1136945834 N * Bertl_oO Bertl
1136945839 M * Bertl back now ...
1136948004 M * duckx 2.6.15 is running fine ..
1136948020 M * duckx Just the firewall did me a headheaq :)
1136948040 M * Bertl aah, good to hear!
1136948058 M * duckx But seems to work fine now :)
1136948064 M * duckx www.mezimail.com
1136948069 M * duckx Running by a vserver :)
1136948106 M * Bertl great! I assume you already added yourself to the 'Happy Users' or 'Providers' page, yes?
1136948117 M * duckx Not yet :)
1136948136 M * duckx But yeah shure we gonna add our self :-)
1136948136 M * Bertl ah, well, if you feel like, go ahead ...
1136948166 M * duckx Well, we are happy users that is pretty shure ...
1136948184 M * duckx It is running for about a year like a charm
1136948212 M * duckx We just did the kernel upgrade ... Well still works fine.... what else could I say ;)
1136948225 M * duckx Thx may be ;)
1136948227 M * Bertl that's the way it's supposed to be ...
1136948261 M * duckx Well we do have an uml box somewhere ...
1136948268 M * Bertl you're welcome! and if you feel like donating time or money, do not hesitate ...
1136948291 M * duckx It just does not work as well ;)
1136948309 M * Bertl UML was really great in the beginning (before xen)
1136948312 M * duckx Always swapping and lagging :=
1136948343 M * duckx UML = no memory efficient :)
1136948370 M * Bertl well, neither is xen, but at least xen distributes the resources better and executes with (almost) native speed
1136948379 M * duckx Yop
1136948387 M * duckx The live cd is pretty impressiv
1136948399 M * Bertl of course, both can not compete with proper resource sharing :)
1136948402 M * duckx I plugged it in a vmware ...
1136948440 M * duckx Well, xen is a great technology, but they don't really the kernel devel ..
1136948461 M * duckx The latest stable is on the 2.6.12 kernel ...
1136948487 M * duckx Damn, it is just like working with my Suse Linux Enterprise Server stuff ;)
1136948497 M * Bertl really? I thought they were aiming for mainline ...
1136948516 M * duckx I did not get that one ...
1136948533 M * duckx What about vserver .. :) In the mainline I mean ? :)
1136948583 M * Bertl well, the question arises every now and then, and although my opinion varies from day to day (it's basically a mood thing :) I do not see an advantage in that (right now)
1136948615 M * duckx Well it is ok for me anyway :)
1136948625 M * Bertl for me personally it's probably more work, and the 'customers' will not benefit that much from it ...
1136948654 M * duckx Well, to me it is pretty simple the way it is ;)
1136948667 M * duckx I just need to know how to use the patch command ;)
1136948681 M * daniel_hozac do you? the step by step guide should have that.
1136948690 M * Bertl lol
1136948711 M * duckx :)
1136948737 M * duckx Man it is just time for me to live and have some nice dreams :)
1136948784 M * duckx Nights !
1136948802 M * Bertl btw, I think that we should try to get a bunch of simple but really useful things implemented and/or assembled
1136948806 M * Bertl duckx: night!
1136948833 M * daniel_hozac such as?
1136948846 M * Bertl daniel_hozac: what do you think (not in terms of implementing or doing that, more in general) of the following:
1136948856 M * daniel_hozac (disk limit support in the tools springs to mind, with automatic saving and restoration)
1136948904 M * Bertl - a pretty simple web frontend (php/perl) which can install a guest, copy a template, set a few paramters start and stop guests (nothing complicated, but something to start a web application from)
1136948946 J * undefined ~undefined@adsl-68-93-109-94.dsl.rcsntx.swbell.net
1136948960 M * Bertl - some scheduler daemon which 'adjusts' scheduling for all guests on the fly, i.e. you set a share or soemthing and all guests get automatically updated every minute or so
1136948963 M * Bertl welcome undefined!
1136948976 M * undefined helloBertl
1136949010 M * Bertl - an admin package of xid aware tools to manipulate stuff
1136949053 M * Bertl - some kind of simple rrd setup for monitoring all kinds of stuff regarding guests and system usage ... no frontend, this can be done by cactis or whatever ...
1136949077 M * daniel_hozac didn't someone promise lots of scripts for monitoring on the mailing list not too long ago?
1136949086 M * daniel_hozac or did i read too much in to that mail...
1136949114 M * Bertl well, folks 'promise' all kinds of stuff, very seldom they keep up to those promises ...
1136949152 M * Bertl I'm more thinking of 'proof of concept' things, which are not perfect or polished on purpose ...
1136949168 M * Bertl something 'other' folks could pick up and 'improve' by using and extending
1136949206 M * daniel_hozac some common infrastructure might be nice for that sort of thing.
1136949242 M * Bertl yeah, I think Hollows library (as well as the util-vserver library) plus a few 'extensions' could be nice
1136949307 M * daniel_hozac i think they all make sense, and they are things people keep asking about on a regular basis.
1136950194 M * Bertl mnemoc: ping? aren't you working with/on T2?
1136950226 M * mnemoc uhm, yes... why?
1136950250 J * stefani ~stefani@c-24-19-46-211.hsd1.wa.comcast.net
1136950267 M * Bertl I'm currently looking into compiling a few things (like bash) with dietlibc, and everytime I google for that, I end up somewhere pointing to Rocklinux or T2 ...
1136950271 M * Bertl welcome stefani!
1136950293 M * mnemoc :p
1136950299 M * Bertl mnemoc: is that coincidence or misinformation?
1136950333 M * mnemoc t2 can build quite complete pure-dietlibc systems
1136950338 M * mnemoc of course including bash
1136950370 M * Bertl excellent! could you point me to some relevant patches for bash and coreutils as well as iproute2?
1136950377 M * mnemoc minute
1136950390 M * Bertl yeah, no problem, take your time ...
1136950398 M * mnemoc http://svn.exactcode.de/t2/trunk/package/base/dietlibc/pkg_patch/ <--- patches
1136950470 M * mnemoc http://svn.exactcode.de/t2/trunk/package/base/dietlibc/parse-config-9 <--- special build oddities (look at the case/esac)
1136950518 M * mnemoc http://svn.exactcode.de/t2/trunk/package/base/dietlibc/pkg_*_post.conf <--- specific actions for some packages, mainly to build static
1136950568 M * mnemoc those files use some helpers and wrappers, but i think it's quite easy to 'realize' what the do
1136950571 M * mnemoc they*
1136950589 M * mnemoc if not, just ask for translation ,-)
1136950605 M * Bertl mnemoc: thanks alot! and another one: "what archs does T2 build on (especially regarding dietlibc)?"
1136950672 M * mnemoc anything supported by gcc
1136950700 M * Bertl hmm .. because I have quite some troubles getting dietlibc to work on more unusual archs
1136950713 M * mnemoc yes
1136950726 M * mnemoc dietlibc is quite incomplete :(
1136950745 M * Bertl are there patches for some archs in T2 (for dietlibc) which are not present upstream?
1136950756 M * mnemoc there is a guy adding ppc64 support and another improving mips support on dietlibc 0.29
1136950778 M * mnemoc look at http://svn.exactcode.de/t2/trunk/package/base/dietlibc/*.patch
1136950796 M * Bertl okay, those are patches _for_ dietlibc, right?
1136950820 M * mnemoc yes, *.patch apply to dietlibc-$ver.tar.bz2
1136950828 M * mnemoc 0.29 in this case
1136950854 M * Bertl i.c. well, thanks again for all the info!
1136950866 M * mnemoc pkg_patch/*.patch are 'remotely' applied to other packages if dietlibc is enabled
1136950957 M * mnemoc crosscompiling -static works very good, diet-dyn and 'native' builds make some troubles because glibc symbols leak in
1136950979 M * Bertl yeah, thought so, but static cross compiling is my main goal
1136951112 M * mnemoc svn co http://svn.exactcode.de/t2/trunk/package/base/dietlibc/ t2-dietlibc    may be useful :)
1136951155 M * Bertl yeah, looks like :)
1136951207 M * mnemoc :)
1136951352 M * Bertl ARCH=ppc diet ppc-linux-gcc -Wall -o hello_world.ppc.diet hello_world.c
1136951360 M * Bertl qemu-ppc hello_world.ppc.diet
1136951360 M * Bertl Hello World!
1136951364 M * Bertl :)
1136951384 M * mnemoc :)
1136953497 P * stefani parting (is such sweet sorrow)
1136953530 J * Smutje ~Smutje@xdsl-87-78-84-170.netcologne.de
1136953636 Q * Smutje_ Ping timeout: 480 seconds
1136956013 P * undefined 
1136956216 J * jpacheco ~justin@CPE00146c1608af-CM0f0099806976.cpe.net.cable.rogers.com
1136956222 M * jpacheco hey everyone
1136956227 M * daniel_hozac hi
1136956250 M * jpacheco has anyone had any luck with running mod_perl 2 in a vserver environment?
1136956285 M * daniel_hozac i haven't tried, but i can't imagine that it would require anything special. what's the problem?
1136956289 M * Bertl jpacheco: why should it give issues?
1136956310 M * jpacheco well im not sure is totally a vserver problem
1136956317 M * jpacheco just wanted to make sure
1136956339 M * jpacheco my mod_perl code is causing apache to respond with Out of memory!
1136956364 M * daniel_hozac memory limits?
1136956395 M * Bertl jpacheco: hmm, could you give a few details to the versions (kernel, patches, utils, apache)
1136956418 M * jpacheco sure
1136956513 M * jpacheco kernel = 2.6.15 patch=2.0.1 distro=gentoo apache=2.0.55
1136956548 M * Bertl okay, do you use memory limits on that specific guest?
1136956596 M * jpacheco no, not that i know of
1136956612 M * jpacheco oh, and the code works fine on another machine
1136956620 M * jpacheco which is not using vserver
1136956629 M * jpacheco however, the software versions differ
1136956736 M * Bertl okay, when and where do you get 'Out of memory' and could you give me some info what the perl code does?
1136956848 M * jpacheco i get the message when i click on a link which brings me to the page
1136956867 M * jpacheco it is that page (the one i am trying to get to) that is giving me the problems
1136956922 M * Bertl hmm, okay ... any ideas what the code for this page actually does?
1136956944 M * jpacheco grabs info from a db
1136956958 M * jpacheco and prints its findings
1136956963 M * jpacheco with graphs
1136956965 M * jpacheco etc
1136957018 M * Bertl could you execute the perl code inside the guest without using apache?
1136957019 M * daniel_hozac do you have any memory limits set in the Apache/mod_perl configuration?
1136957047 M * jpacheco where would that be set?
1136957075 M * daniel_hozac i don't know, PHP has a maximum memory setting in its configuration file.
1136957172 M * jpacheco could this be a CAPs issue? or would CAPs help me out at all?
1136957238 M * Bertl could be, but as long as we do not know _what_ causes the OoM, we are left in the dark
1136958066 M * Bertl okay, I'm off to bed now ... good luck ... 
1136958083 N * Bertl Bertl_zZ
1136958505 J * coocoon ~coocoon@p54A06533.dip.t-dialin.net
1136958523 M * coocoon good morning
1136958530 M * jpacheco do you guys know how to debug in apache2?
1136958536 M * jpacheco for mod_perl 2
1136958555 M * coocoon me not i am sorry
1136958574 M * coocoon have had a look at the vserver mailingvlist
1136959196 Q * jpacheco Quit: BitchX: No windows left!
1136961180 J * |coocoon| ~coocoon@p54A06533.dip.t-dialin.net
1136961335 Q * coocoon Ping timeout: 480 seconds
1136961419 J * balbir ~balbir@59.145.136.1
1136966728 J * NikDaPhreak ~NikDaPhre@217.75.141.95
1136966732 M * NikDaPhreak hi all
1136967005 Q * |coocoon| Ping timeout: 480 seconds
1136967733 Q * shedi Quit: Leaving
1136968241 Q * balbir Quit: Leaving
1136969699 J * prae ~prae@ezoffice.mandriva.com
1136970326 J * meandtheshell ~markus@85-125-229-57.dynamic.xdsl-line.inode.at
1136970972 J * shedi ~siggi@tolvudeild-201.lhi.is
1136971412 Q * duckx Quit: Client exiting
1136971569 Q * zobel Ping timeout: 480 seconds
1136971876 J * zobel zobel@zobel.irc.ftbfs.de
1136972263 J * Mandolin ~kN@office.2forces.net
1136972279 Q * Mandolin Quit: 
1136973129 J * coocoon ~coocoon@p54A0646B.dip.t-dialin.net
1136973641 J * balbir ~balbir@59.145.136.1
1136973989 M * aba hm, would it be possible to allow inside the vserver to only mount proc to all different places, and also to create only certain specific file nodes (like /dev/null)?
1136974799 Q * zobel Ping timeout: 480 seconds
1136975142 M * NikDaPhreak aba: AFAIK - no. You can generally allow or disallow mounts and/or device-creations
1136975169 J * zobel zobel@zobel.irc.ftbfs.de
1136975189 M * aba NikDaPhreak: well, that's what I wrote "would it be possible to allow" and not "how do I" :)
1136975206 M * aba basically, it was "Bertl, can't that be changed" ...
1136975813 J * DavidS ~david@vpn.uni-ak.ac.at
1136975856 M * NikDaPhreak aba: ok :-) 
1136976403 M * DavidS hi, is there an ETA on a patch-2.6.15-vs2.1.0 ? Or am I just blind and didn't find it yet?
1136976531 M * NikDaPhreak DavidS: http://vserver.13thfloor.at/Experimental/patch-2.6.15-vs2.1.0.4.diff
1136976552 M * NikDaPhreak DavidS: http://vserver.13thfloor.at/Experimental/ for older ones
1136976580 M * DavidS NikDaPhreak: are they linked from anywhere or do you just "need-to-know"?
1136976738 M * NikDaPhreak DavidS: well... it is "need-to-know" more or less...
1136976805 M * DavidS NikDaPhreak: is there any reason this link isn't in the wiki?
1136976898 M * NikDaPhreak DavidS: AFAIK - no
1136977053 M * DavidS NikDaPhreak: good, so I put a link on the wiki ...
1136977120 M * NikDaPhreak ok
1136977222 M * Aiken that link is on the wiki
1136977415 M * DavidS Aiken: :( now that you say it ...
1136977441 M * Aiken [Bleeding Edge Stuff] (prereleases, release candidates)
1136977492 M * DavidS I didn't expect it under "Important Links" seperatly from "Downloads"
1136977595 M * DavidS any ideas how the general situation on the wiki could be improved? again and again, i get lost :(
1136977645 M * DavidS perhaps split distribution specific things into their own sections, leaving generic development/release stuff in a less crowded center?
1136977919 J * Doener doener@i5387D636.versanet.de
1136977995 Q * coocoon Ping timeout: 480 seconds
1136978486 Q * _are_ Ping timeout: 480 seconds
1136978660 J * coocoon ~coocoon@p54A07158.dip.t-dialin.net
1136978851 Q * Aiken Ping timeout: 480 seconds
1136979346 P * DavidS 
1136979379 J * undefined ~undefined@adsl-68-93-109-94.dsl.rcsntx.swbell.net
1136980093 J * id23 ~id@p54A02BA4.dip0.t-ipconnect.de
1136980116 M * id23 lo
1136981946 J * dothebart ~willi@xdsl-213-196-221-115.netcologne.de
1136982389 Q * tudenbart Ping timeout: 480 seconds
1136983062 Q * balbir Quit: Leaving
1136983737 J * User08 ~User08@0x5551697e.adsl.cybercity.dk
1136983922 M * Hollow morning
1136983941 M * coocoon hello
1136983975 M * lchvdlch morning
1136984080 M * Hollow NikDaPhreak: btw, 1.0.2 is released.. did you have chance to test the network things?
1136984130 M * coocoon hollow:   please can u tell me is it advisable to create a virtual lo and route the virtual servers through the virtual eth0:0 for example 192.168.1.0
1136984141 J * oliwel ~oliwel@ldvpc07.ldv.e-technik.tu-muenchen.de
1136984145 A * oliwel waves hello to the crowd
1136984149 M * Hollow you cannot create a "virtual lo"
1136984170 M * NikDaPhreak Hollow: unfortunately - no. had to take care of a crashed server :-( "Bad Linux, Bad Linux!" ;-)
1136984180 M * Hollow is it running gentoo?
1136984182 M * Hollow :P
1136984209 M * coocoon hollow: as i am reading here"Linux wegweiser fuer Netzwerker" it is possible
1136984218 M * Hollow coocoon: to be exact.. you cannot create any virtual device atm
1136984229 M * NikDaPhreak Hollow: err... I'm Slackware fan for ages... ;-)
1136984230 M * coocoon ok configure
1136984259 M * Hollow coocoon: the networking is completely done on the host, you can just give context the permission to only use a certain subset of ip adresses
1136984305 M * Hollow NikDaPhreak: too bad ;)
1136984308 N * User08 kongsted
1136984319 M * kongsted Hi
1136984333 M * kongsted I have some problems under the installation
1136984363 M * kongsted I use CentOs 4.2 - Kernel 2.6.14.6 - VS 2.01
1136984368 M * coocoon I think I must route the vip's
1136984394 M * kongsted I had followed Step-by-Step+Guide+2.6
1136984403 M * Hollow coocoon: to route from and to where?
1136984431 M * coocoon hollow: from my configured network o the host
1136984437 M * coocoon to the net
1136984447 M * kongsted My first problem: /etc/init.d/vprocunhide restart
1136984455 M * coocoon hollow: mustn't i do this 
1136984459 M * Hollow i assume your network access is configured properly on the host?
1136984468 M * coocoon yes
1136984468 M * kongsted I can't find anything on vprocunhode
1136984478 M * Hollow do you have one or more public ips?
1136984484 M * coocoon one
1136984495 M * Hollow ok, you're done then ;)
1136984526 M * coocoon I imagine I have a LAN
1136984540 M * coocoon hollow: so I must configure a router mustn't I
1136984563 M * coocoon to get my vservers accessible from outside
1136984569 M * Roey kongsted:  2.6.14.6 is out??
1136984570 M * Hollow no, that's why i said networking is done on the host... 
1136984577 M * Roey kongsted:  I thought .5 was the latest?
1136984586 M * Hollow you don't have virtual network devices or virtual iptables or sth like this
1136984589 M * kongsted Roey: I guess so... :-)
1136984593 M * Roey heh
1136984593 M * Roey thanks
1136984605 M * Hollow you configure the IPs on the host, and give your vserver only access to a subset of them
1136984613 M * kongsted http://www.kernel.org/pub/linux/kernel/v2.6/linux-2.6.14.6.tar.gz
1136984660 M * coocoon yes but how to access them from outside if they have normal uips like 192.168....
1136984672 M * coocoon configuring dns 
1136984676 M * coocoon or what
1136984802 M * kongsted And if I try testme.sh, I got the following error:
1136984807 M * kongsted Linux-VServer Test [V0.14] Copyright (C) 2003-2005 H.Poetzl
1136984814 M * kongsted Can't set the new security context
1136984821 M * kongsted : Invalid argument
1136984827 M * kongsted chcontext failed!
1136984831 M * kongsted Can't set the ipv4 root (Invalid argument)
1136984835 M * kongsted chbind failed!
1136984840 M * kongsted Linux 2.6.14.6 x86_64/0.30/0.30 [E] (0)
1136984845 M * kongsted VCI:  0002:0001 236 03000016
1136984892 M * kongsted What did I forget?
1136985048 M * coocoon kongsted: http://linux-vserver.org/CentOS_HowTo
1136985061 M * coocoon kongsted: maybe u will find here the answer
1136985082 M * schellh well i had this error because i used a gentoo vserver howto
1136985094 M * schellh it was gone when i compiled with the vserver api and with dynamic contexts
1136985140 M * kongsted coocoon> Thanks... I try that...
1136985150 Q * kongsted Quit: kongsted
1136985228 P * undefined 
1136985482 M * Doener schellh: actually his kernel version doesn't look like it was patched and the tools are ancient ;)
1136985579 M * Doener hm, unpatched kernel wouldn't have VCI though...
1136985716 M * Hollow coocoon: port forwarding, nat..
1136985940 M * coocoon hollow: yes
1136985959 M * coocoon hollow: this is what I tried => iptables -t nat -A PREROUTING -p tcp --destination-port 10022 -j DNAT --to-destination 192.168.1.110
1136985970 M * Hollow you also need SNAT
1136985984 M * Hollow else your guest can't send packets to the inetnet
1136986015 M * coocoon hollow: => /sbin/iptables -A POSTROUTING -t nat -s 192.168.1.110 -o "eth0" -j SNAT --to-source "hostIP"
1136986041 M * coocoon these both commands I have configuired in pre-start.d
1136986044 M * coocoon or set
1136986045 M * coocoon sorry
1136986087 M * Hollow yeah, should be fine then..
1136986092 J * Milf ~Miranda@ipsio388.ipsi.fraunhofer.de
1136986099 M * Hollow don't forget /etc/resolv.conf inside your guest
1136986115 M * coocoon hollow:  from my host resolv.conf
1136986126 M * coocoon hollow: ?
1136986138 M * Hollow yeah
1136986155 M * Roey hey does Intel have any Virtualization Technology-supporting CPUs out yet?
1136986167 M * coocoon hollow: i mustn't configure dns
1136986170 M * Roey http://www.intel.com/technology/computing/vptech/  says yes
1136986173 M * Roey like
1136986181 M * coocoon hollow: for that
1136986182 M * Roey does the Xeon we just got support VT??
1136986203 M * coocoon hollow: to reach my vserver like vhost00....net
1136986239 M * Hollow coocoon: you have to configure dns to point to your public ip
1136986253 M * coocoon hollow: resol.conf is configured as on my host
1136986298 M * coocoon hollow: I can reach my host with ping
1136986384 M * coocoon hollow: ping vhost00.....net no packages will send
1136986447 M * Hollow coocoon: and vhost00 resolves to?
1136986545 M * coocoon hollow: no
1136986614 M * coocoon hollow: I haven't configured till yes therefore I am asking
1136986629 M * coocoon hollow: yesterady I have had it but it was the same
1136986645 M * coocoon ssh client stays for a short time and then stopped
1136986656 M * coocoon hollow: my ssh cleint
1136987038 M * coocoon hollow: on my host it works
1136987048 M * coocoon hollow: ping vhost00
1136987051 M * coocoon but from outside
1136987056 M * coocoon not
1136988476 J * undefined ~undefined@adsl-68-93-109-94.dsl.rcsntx.swbell.net
1136990055 J * mef ~mef@targe.CS.Princeton.EDU
1136990059 Q * mef Quit: 
1136991427 J * Lars ~larspower@you.are.the.lamer.la
1136991661 M * Lars Does http://www.13thfloor.at/vserver/s_rel26/v2.01/patch-2.6.14.3-vs2.01.diff work for linux-image-2.6.14-2-686 also?
1136992087 N * Bertl_zZ Bertl
1136992106 M * Bertl Lars: what _is_ linux-image-2.6.14-2-686 ?
1136992114 M * Bertl morning folks!
1136992153 M * NikDaPhreak Bertl: :-)
1136992375 M * Loki|muh Bertl: i would guess a debian-package 
1136992518 M * aba morning, Bertl 
1136992587 M * Bertl Loki|muh: hmm, well, could also be anything else, but somehow looks binary to me :)
1136992659 M * SiD3WiNDR debian stuff is kernel-image-blah
1136992665 M * SiD3WiNDR and those are indeed binary
1136992866 M * Bertl ah, well, I guess Lars will return sooner or later :)
1136992911 M * Lars Yay
1136992938 M * Lars Is there a .diff file for kernel 2.6.14-2-686 ?
1136992998 M * undefined SiD3WiNDR: debian stuff *use* to be "kernel-image-*" but now that debian is adding support for bsd (which has its own kernel), the package name had to be more specific, ie linux-kernel-*
1136993000 M * Lars nm ill use -3-
1136993020 M * undefined argh!
1136993027 M * undefined linux-image-*
1136993081 M * SiD3WiNDR undefined: :o
1136993097 M * SiD3WiNDR undefined: I see
1136993213 Q * coocoon Quit: KVIrc 3.2.0 'Realia'
1136993304 M * undefined Lars: try kernel-patch-vserver (= 2.01) as it suggests linux-source-2.6.14 (see http://packages.debian.org/unstable/devel/kernel-patch-vserver)
1136993352 Q * oliwel Remote host closed the connection
1136993364 M * Lars i got the kernel 2.6.14-2 installed from linux-image i copied the 2.6.14-2 config file from /boot into my kernel source 2.6.14-3 then i got the .diff patch 2.6.14-3 and patched the kernel. Now im compiling the kernel
1136993368 M * Lars it should work right?
1136993386 M * micah Lars: are you using the diff from kernel-patch-vserver? 
1136993387 M * undefined did the patch apply cleanly?
1136993404 M * Lars undefined: i got the patch from http://www.13thfloor.at/vserver/s_rel26/v2.01/patch-2.6.14.3-vs2.01.diff
1136993412 M * Lars and yes the patch applyed cleanly.
1136993413 M * micah Lars: and where does the kernel source 2.6.14-3 come from?
1136993415 M * undefined Lars: micah is here, he can answer all your questions ;)
1136993423 M * Lars micah: www.kernel.org
1136993439 M * micah sure that will work
1136993445 M * Bertl Lars: kernel org now has debian packages?
1136993447 M * micah it wont be the debian kernel
1136993456 M * Lars Bertl: No.
1136993468 M * micah I suspect that he actually just has 2.6.14 kernel source, not 2.6.14-3
1136993470 M * Bertl Lars: I guess your numbers are a little off, let me clarify
1136993489 M * Bertl Lars: from kernel org, you want to get 2.6.14.6 or 2.6.15
1136993503 M * Bertl Lars: there is no 2.6.14-3 on kernel org
1136993507 M * Bertl (IMHO)
1136993519 M * micah your humble opinon is right :)
1136993527 M * Lars oh it was linux-2.6.14 bla bla i got.
1136993535 M * Lars .3
1136993556 M * Bertl good, get the 2.6.14.6 in this case, only contains fixes and the patches for .3 apply 
1136993559 M * undefined actually, isn't there a debian mirror on kernel.org that would have the debian linux kernel packages ;)
1136993572 M * aba undefined: it is, yes
1136993579 M * micah Bertl: do you recall if this person ever come on IRC and get this issue resolved: http://list.linux-vserver.org/archive/vserver/msg10682.html ?
1136993580 M * aba even one of the top-level mirrors
1136993592 M * micah Bertl: I've got someone asking me about the same issue
1136993602 M * Lars Soo. wget http://www.kernel.org/pub/linux/kernel/v2.6/linux-2.6.15.tar.bz2 - unzip it than cp /boot/config-2.6.14-2-686 /usr/src/linux-2.6.16.tar.bz2 then cd to the dir and patch it with the diff
1136993603 M * Lars right?
1136993637 M * Bertl no, sounds very strange ...
1136993638 M * micah Lars: should work
1136993665 M * Bertl what is 2.6.16 now? and why copy the config over it?
1136993667 M * micah Lars: however, you would want to cp /boot/config-2.6.14-2-686 /usr/src/linux-2.6.15/.config
1136993675 M * micah and then run make oldconfig
1136993676 M * Lars yes sorry.
1136993693 M * undefined Lars: be aware that there may be patches in the debian kernel (not in the upstream kernel) that are reflected in the config file
1136993707 M * Lars Hmm
1136993713 M * Lars Well
1136993713 M * micah undefined: i think make oldconfig will resolve that
1136993716 M * Lars what would be the best way than?
1136993725 M * undefined Lars: so just because something is in the original config file you are copying over, you won't have those features in the compiled kernel
1136993775 M * micah Lars: you make a choice between having the debian kernel or the vanilla one
1136993789 M * Lars What's the best?
1136993797 M * micah that is a subjective question
1136993799 M * Bertl micah: IIRC, this was resolved by two things, first, extending the 'kill' syscall command to send the signal to all processes including init, an then by a separate 'kill on reboot' flag - the tools 0.30.209 probably only use the first one, but it should be sufficient ...
1136993848 M * Lars micah: Soo if i should do it the debian way how should it go than? apt-get install linux-image-2.6.15-1-686 and than?
1136993889 M * micah Bertl: was the 'kill' syscall extention done in 2.0.1 by any chance?
1136993905 M * Bertl micah: but, if the problem is that the 'guest' is not terminating properly, then I have no idea why and how this was solved, IIRC again, Hollow fixed the gentoo guests around that time
1136993922 M * Bertl micah: it is in 2.0.1
1136993945 M * micah Lars: if you want to do it the debian way, then you can follow this: http://deb.riseup.net/vserver/preparing/
1136993964 M * micah it only details 2.6.8 and 2.6.12 at the moment, but its the same procedure
1136993995 M * micah Bertl: ah, ok, this person is running a -rc of 2.0 so upgrading their kernel patch will probably make it go away
1136994011 M * Lars micah: but in apt-get there are only 2.6.8 kernel sources.
1136994017 M * Lars soo i would have to get one from kernel.org
1136994060 M * micah Lars: if there are only 2.6.8 kernel sources, then you are using stable
1136994082 M * micah Lars: and you aren't looking carefully, because there are also 2.4.27 kernel sources in stable :)
1136994085 M * Lars no, im running unstable.
1136994110 M * micah Lars: there are no kernel sources for 2.6.8 in debian unstable
1136994120 M * micah Lars: ah, you may be looking for the wrong package name
1136994126 M * micah Lars: the name changed
1136994138 M * Lars linux-source ..
1136994148 M * micah right, as that page says
1136994201 M * micah Lars: there is work going on to prepare a kernel image package with vserver enabled, should be coming soon
1136994260 M * Bertl micah: I would assume so ...
1136994316 M * Lars But ill just do it the other way: wget http://www.kernel.org/pub/linux/kernel/v2.6/linux-2.6.15.tar.bz2; tar jxvf linux-2.6.15.tar.bz2; cd linux-2.6.15; cp /boot/config-2.6.14-2-686 /usr/src/linux-2.6.15/.config; wget http://www.13thfloor.at/vserver/s_rel26/v2.01/patch-2.6.14.3-vs2.01.diff; cat patch-2.6.14.3-vs2.01.diff | patch -p1 and then make oldconfig and... make-kpkg --rootcmd fakeroot --revision custom01 --added-patches vserver --append-to-ver
1136994351 M * waldi hmm, can I kill processes from xid 1?
1136994351 M * Bertl micah: another relevant fix could be the signal code fix in 2.0-rc3 (if your friend uses and older rc)
1136994376 M * Lars Noo what woudnt work. Im getting failed when i patch
1136994376 M * Lars hmmm
1136994402 M * Bertl waldi: no, but from xid=0
1136994412 M * Lars micah: why is the patching failing?
1136994429 M * micah Lars: I dont think you can use make-kpkg with vanilla kernel sources, but I could be wrong
1136994430 M * waldi Bertl: i can't see them but kill them?
1136994449 M * Bertl Lars: it is failing because you need a different patch for 2.6.15
1136994453 M * micah Lars: perhaps the patch-2.6.14.3-vs2.01.diff doesn't apply to 2.6.15?
1136994456 M * Lars Bertl: which one
1136994479 M * Bertl waldi: yes, in two ways, either 'just' by the pid for now, or (the right way to do it) with the vkill command
1136994490 M * micah Bertl: hm it turns out he is using the 2.6.8 patch, with 0.30.209, so it must be something that util-vserver changed
1136994493 M * waldi Bertl: hmm, does not work
1136994512 M * Bertl Lars: http://vserver.13thfloor.at/Experimental/patch-2.6.15-vs2.0.1.2.diff
1136994520 M * waldi Bertl: i started the vserver with the --rescue option
1136994534 M * Bertl waldi: aha, okay ...
1136994545 M * waldi and I can't kill the sleep process
1136994614 M * Bertl chcontext --xid 42 sleep 100 &
1136994627 M * Bertl vps auxwww | grep sleep
1136994638 M * Bertl root      1860    42                0.0  0.0   6652   616 pts/0    S    17:39   0:00 sleep 100
1136994645 M * Bertl vkill --xid 42 -s 9 -- 1860
1136994694 M * waldi # vps aux | grep sleep   
1136994694 M * waldi root     16022 49164 database       0.0  0.0  2744  584 ?        Ss   15:39   0:00 /bin/sleep 900
1136994698 M * waldi # vkill -c 49164 -s 9 -- 16022
1136994702 M * waldi # vps aux | grep sleep        
1136994703 M * waldi root     16022 49164 database       0.0  0.0  2744  584 ?        Ss   15:39   0:00 /bin/sleep 900
1136994926 M * Lars micah: I now did make oldconfig now im running make and then ill run make modules_install and then ill cp .config /boot/config-2.6.15-vs and cp System.map to /boot/System.map-2.6.15-vs and then cp arch/i386/boot/bzImage /boot/vmlinuz-2.6.15-vs and than mkinitrd -o /boot/initrd.img-2.6.15-vs 2.6.15-cs
1136994944 M * Bertl waldi: try to send a CONT signal
1136994961 M * waldi Bertl: too late, it exited after the 900 seconds
1136994966 M * Bertl waldi: and start avoiding dynamic contexts :)
1136994976 M * waldi Bertl: why?
1136994984 M * Lars micah: I suppose this will work.
1136994989 M * Bertl waldi: because they will go away soon ...
1136995005 M * waldi okay, that is a reason
1136995031 M * Bertl well, maybe userspace will provide this feature in the future
1136995045 M * Bertl but at least the kernel will get rid of dynamic contexts
1136995160 M * Bertl Lars: well, the System.map-* only makes sense if it matches the kernel version
1136995200 M * Lars Bertl: it will match the kernel im compiling yes?
1136995222 M * Bertl in your case I assume that would be /boot/System.map-2.6.15-vs2.0.1.2
1136995240 M * waldi Bertl: no I have to find how to get ride of context mounts
1136995262 M * Bertl waldi: hmm?
1136995292 M * waldi Bertl: hmm, something is wrong here
1136995309 M * waldi # rm -r *
1136995309 M * waldi zsh: sure you want to delete all the files in /srv/vservers/database [yn]? y
1136995309 M * waldi rm: cannot remove directory `dev/pts': Device or resource busy
1136995309 M * waldi rm: cannot remove directory `proc': Device or resource busy
1136995309 M * waldi rm: cannot remove directory `tmp': Device or resource busy
1136995320 M * waldi but now it works
1136995342 M * Bertl well, mounts don't go away before the last process exits
1136995425 M * Bertl (the last process using the namespace, that is)
1136995643 M * Bertl okay, leaving now ... back later this evening ...
1136995649 M * Bertl have fun!
1136995654 N * Bertl Bertl_oO
1136995946 N * nokoya Guest49296
1136996074 N * Guest49296 nokoya
1136996152 Q * shedi Quit: Leaving
1136996848 Q * Milf Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org
1136996979 P * undefined 
1136997030 Q * mcp Read error: Connection reset by peer
1136998034 M * kilian Bertl_oO: uhm, i guess my hppa kernel is now working almost..
1136998037 M * kilian Bertl_oO: what's "[011]# failed.
1136998037 M * kilian [011]# failed.
1136998043 M * kilian telling me?
1136998051 M * kilian uhm, once only =)
1136998122 J * mcp ~hightower@wolk-project.de
1136998286 M * micah kilian: that is a security check regarding nodes
1136998299 M * micah kilian: do you have linux-capabilities configured as modules?
1136998313 M * micah and if so, is it loaded? (lsmod)
1136998345 M * micah modprobe capability, then remove any device nodes created in /tmp and re-do testme.sh
1136998369 M * kilian micah: FATAL: Module capability not found.
1136998387 M * kilian micah: it's built in as:
1136998387 M * kilian CONFIG_SECURITY_CAPABILITIES=y
1136998459 M * micah kilian: which patch is it? the 2.1 or 2.01?
1136998464 M * kilian 2.1.0.3
1136998579 M * micah hmm, every time I've seen 011 fail it was due to the capability system being broken (can be in the userspace)
1136998755 M * kilian micah: well, you can have my vserver-kernel-image if you wanna try yourself =)
1136998769 M * kilian micah: trick was to use klibc 1.1.16 instead of 1.1.14 ;)
1136999353 J * Smutje_ ~Smutje@xdsl-87-78-17-121.netcologne.de
1136999461 Q * Smutje Ping timeout: 480 seconds
1136999498 Q * NikDaPhreak Quit: bbl...
1136999629 J * fosco fosco@tao.mu
1136999632 M * fosco hi
1136999661 M * fosco I have some problems with vdlimit
1136999705 M * fosco limit seems to be set (according to df), but I can write a lot more than the limit, including when df reports that 0 bytes are available :\
1136999719 J * stefani ~stefani@superquan.apl.washington.edu
1137000377 M * fosco shame on me
1137000380 M * fosco tagxid :)
1137000432 J * Viper0482 ~Viper0482@p54975E25.dip.t-dialin.net
1137000614 Q * zobel Read error: Connection reset by peer
1137000625 J * zobel zobel@zobel.irc.ftbfs.de
1137000856 J * menomc ~amery@200.75.27.102
1137000914 J * undefined ~undefined@adsl-68-93-109-94.dsl.rcsntx.swbell.net
1137000961 Q * mnemoc Ping timeout: 480 seconds
1137000961 N * menomc mnemoc
1137001163 J * Arie- Arie@dsl-82-199-145-174.dutchweb.nl
1137001167 P * Arie- 
1137001415 Q * Viper0482 Remote host closed the connection
1137002779 P * undefined 
1137003893 Q * prae Quit: Execute Order 69 !
1137004092 M * brc -otagxid
1137004102 Q * Doener Quit: Leaving
1137004107 J * Doener doener@i5387D636.versanet.de
1137004435 J * NikDaPhreak ~NikDaPhre@193.24.241.34
1137004982 Q * Doener Ping timeout: 480 seconds
1137005013 J * Doener doener@i5387D409.versanet.de
1137005614 J * undefined ~undefined@adsl-68-93-109-94.dsl.rcsntx.swbell.net
1137006030 Q * Hollow Read error: Connection reset by peer
1137006051 J * Hollow ~hollow@home.xnull.de
1137007113 P * undefined 
1137007125 J * undefined ~undefined@adsl-68-93-109-94.dsl.rcsntx.swbell.net
1137008571 J * webmester ~webmaster@0x5551697e.adsl.cybercity.dk
1137008590 N * webmester kongsted
1137008594 M * kongsted Hi
1137008604 M * kongsted Still have some problems...
1137008622 M * kongsted I trying to install vserver on a Centos 4.2
1137008639 M * kongsted I had compiled a new kernel - correct I guess...
1137008675 M * kongsted Almost standard setup - is it possible that I missed something during the compiling?
1137008730 M * kongsted My problem is that it don't recognize my 64-bit cpu 
1137008752 M * brc did you choose the right processor during kernel config ?
1137008771 M * kongsted as far as I remember - yes....
1137008814 M * kongsted But if it is the only reason that it doesn't find my cpu - it must be the reason...
1137008828 M * daniel_hozac it doesn't find your CPU?
1137008847 M * daniel_hozac what on earth makes you think that?
1137008875 M * kongsted after I install the util I can't find anything in /usr/lib64 with vserver....
1137008892 M * kongsted But there is some files in /usr/lib...
1137008903 M * daniel_hozac inside the vserver?
1137008909 M * kongsted ?
1137008911 M * daniel_hozac or the utils themselves?
1137008919 M * kongsted No the util...
1137008933 M * daniel_hozac well, that'd be due to how you configured the utils.
1137008949 M * kongsted I haven't complete the install because of that...
1137008973 M * kongsted I followed http://linux-vserver.org/CentOS_HowTo
1137008996 M * daniel_hozac just add --libdir=/usr/lib64
1137009032 M * kongsted First http://linux-vserver.org/Step-by-Step+Guide+2.6 and afterwards http://linux-vserver.org/CentOS_HowTo
1137009050 M * kongsted thanks - I will try that...
1137009052 M * kongsted :-)
1137009078 M * daniel_hozac shouldn't really matter though unless you intend on installing a 32-bit version as well.
1137009093 J * shedi ~siggi@inferno.lhi.is
1137009224 M * kongsted Well.... Now I have some data in /usr/lib64.... But not ./distribution.... :-(
1137009275 M * daniel_hozac did you remember make install-distribution?
1137009286 M * kongsted no...
1137009299 M * kongsted It is first time I see that command...
1137009305 M * kongsted (or option)
1137009336 M * kongsted make: *** No rule to make target `install-distribution'.  Stop.
1137009408 M * daniel_hozac what utils are you installing, exactly?
1137009425 M * kongsted util-vserver-0.30
1137009437 M * kongsted util-vserver-0.30.tar.bz2
1137009449 M * daniel_hozac 0.30? why not 0.30.209?
1137009487 M * kongsted hmmmm..... Great question... :-)
1137009496 M * kongsted I will try that... :-)
1137009535 M * daniel_hozac as is suggested in the step by step guide.
1137009559 M * NikDaPhreak Hollow: ping
1137009596 M * kongsted daniel_hozac Sorry....
1137009630 A * kongsted sit down on my chair again and look like a mouse.... :-)
1137010001 N * Bertl_oO Bertl
1137010006 M * Bertl evening folks!
1137010044 M * Roey ciao!
1137010163 M * Bertl hey mcp! LTNS!
1137010183 M * Bertl kilian: so how was the klibc?? related?
1137010192 M * Bertl s/the/that/
1137010208 A * kongsted say godnight...
1137010214 Q * kongsted Quit: 
1137010469 M * brc bertl
1137010469 M * brc !
1137010497 M * brc Bertl do you have plans  make swap with memory limits available ?
1137010512 M * brc i was checking freebsd and they do that, why not getting some "help" from their code ? :)
1137010642 M * meandtheshell brc: what does this mean exactly "swap with memory limits"?   limiting the amount of swap a guest could use ... hm? makes no sense
1137010650 M * Bertl well, no problem with that, but a) what precisely do you have in mind?
1137010657 Q * lilalinux Remote host closed the connection
1137010657 M * kilian Bertl: yep
1137010657 M * kilian Bertl: it was the 1.1.14 being broken
1137010683 M * Bertl kilian: pardon the question, but what is klibc, and what has it to do with the kernel?
1137010711 M * brc meandtheshell: If you have a guest with memory limits, it won't swap. When the limit is reached it kills processes that try to fork()
1137010717 M * kilian Bertl: it's initramfs tools which need to be run in the scripts like to find the root-fs-type
1137010756 M * Bertl kilian: okay, and again, how is this related to the capability system being disabled?
1137010765 M * kilian Bertl: http://www.kernel.org/pub/linux/libs/klibc/
1137010773 M * kilian Bertl: that one isn't related
1137010779 M * kilian Bertl: that's just "another" problem ;)
1137010787 M * kilian Bertl: the first part was getting the thing booting..
1137010807 M * Bertl kilian: ah, okay, fine, I got the impression that this was the solution for the 011 failure
1137010808 M * kilian Bertl: now it's booting ok and running, but there's the results i've given above..
1137010818 M * kilian Bertl: nope, that was the solution to *GETTING THERE* ;)
1137010844 M * kilian Bertl: as in "hey, we have new errors, we're progressing" ;)
1137010846 M * Bertl okay, you have the kernle installed somewhere, I guess, should I have a look at it (userspace side)?
1137010864 M * Bertl brc: consider the following situation:
1137010867 M * kilian Bertl: whatever you wish, you can also try to talk me into it
1137010898 M * Bertl brc: you have 10 guests, all limited to let's say 64MB of 'ram' and the same amount of 'swap'
1137010941 M * Bertl brc: and all of them are 'slightly' above the 64MB limit, let's say 80MB, what do you think will the box (the host) with 2GB memory do most of the time?
1137011017 M * Bertl kilian: well, you ahve my key, so just let me know ...
1137011025 M * kilian Bertl: you have ipv6?
1137011038 M * Bertl kilian: nope
1137011044 M * kilian Bertl: *sigh* ok.. hold on
1137011067 M * brc Bertl: it would be better if the host would use physical RAM not swap
1137011152 M * brc Bertl: Until ram is full. Then it would SWAP the guests which have exceeded their limits. 
1137011232 M * kilian Bertl: mmh, your ssh-pubkey was on the test-ppc64 host, right?
1137011258 M * kilian Bertl: i guess you'd rather get me another copy, as waldi is still busy AFK
1137011286 M * Bertl brc: see, that's the idea, and that is why we will do 'soft' limits for memory pretty soon, but for now they are 'hard' limits. period.
1137011307 M * Bertl kilian: np, sec
1137011362 M * FaUl Bertl: going to visit that ultrasparcs tomorrow and just ordered the plugs necessary for that serial-consoles right now, hope they will ship soon
1137011409 M * brc Bertl: ok, that will be great :)
1137011464 M * Bertl brc: for now, the 'suggestion' is to use roughly half the available memory for the hard limit
1137011484 M * Bertl brc: that will keep folks from hogging memory, but will not affect 'friendly' users
1137011498 M * Bertl brc: (or double the 'typical' amount)
1137011513 M * meandtheshell Bertl: that soft limit at RAM usage - will it work similar like the bucket method for CPU usage per guest
1137011528 M * meandtheshell ?
1137011545 M * FaUl and i'm negoating for a faster internet-uplink, too, that will speed up ssh-connections :-)
1137011596 M * Bertl meandtheshell: no, the idea is to do the following things:
1137011626 M * Bertl - leave memory in ram when possible
1137011640 M * Bertl - when swapping, prefer guests which are over their limit (soft)
1137011662 M * Bertl - add some penalty (scheduling) for guests over their limit (on page access)
1137011955 M * meandtheshell ok - sound's smart - so: 
1137011958 M * meandtheshell soft limit = (physical RAM - constant amount which is needed by host etc.) / number of guests
1137012072 M * meandtheshell but it could of course be differnt for different guests
1137012083 M * Bertl yes, that's the idea ...
1137012709 M * Hollow NikDaPhreak: pong
1137012745 M * Bertl FaUl: excellent, so it will get a console and maybe the 9GB disk?
1137013343 M * FaUl Bertl: yes, indeed
1137013396 M * Bertl great! really great! TIA!
1137013584 Q * Lars Read error: Connection reset by peer
1137014948 M * NikDaPhreak Hollow: I dot an error message while trying to compile vserver-utils-1.0.2
1137014999 M * NikDaPhreak Hollow: if I install the util-vserver-0.30.209 - a bunch of error messages and if I remove the util-vserver package - nidof.c:28:21: vserver.h: No such file or directory
1137015016 M * Hollow NikDaPhreak: you need to install libvserver before
1137015030 M * Hollow http://dev.croup.de/proj/libvserver
1137015034 M * NikDaPhreak Hollow: err... ok, I missed that ;-)
1137015062 M * NikDaPhreak Hollow: my fault... I might need a bit more sleep these days ;-)
1137015068 M * Hollow i splitted out the syscall wrapper stuff to its own lib
1137015082 M * Hollow no problem with insomnia.. ;)
1137015248 M * NikDaPhreak Hollow: well... some days ago I installed this with no probs, but now I torchured some other server ;-) 10x for the tip :-)
1137015301 M * Hollow you're welcome!
1137015373 M * Hollow maybe i should add some check to vserver-utils for autoconf
1137015428 A * NikDaPhreak gone sleeping....
1137015436 M * Bertl nigh NikDaPhreak!
1137015519 M * Hollow heya Bertl
1137015774 J * Viper0482 ~Viper0482@p549767E5.dip.t-dialin.net
1137015840 M * Bertl welcome Viper0482!
1137016705 Q * Viper0482 Read error: Connection reset by peer
1137016758 J * Aiken ~james@tooax6-017.dialup.optusnet.com.au
1137016814 M * Bertl welcome Aiken!
1137016829 J * Viper0482 ~Viper0482@p549767E5.dip.t-dialin.net
1137016857 M * Aiken hello
1137016895 M * Aiken noticed 2.1.0.X went through a few releases yesterday  .1 to .4
1137016903 M * Aiken tidying up or bug fixes?
1137016915 Q * tso Quit: Connection reset by pheer
1137017008 M * Aiken maybe at least some tidying up, 2.1.0.4 applies cleaner than .1 did
1137017710 M * SiD3WiNDR Usage: /usr/sbin/vserver {start|stop|suexec|restart|condrestart|exec|enter|chkconfig|running|status}
1137017714 M * SiD3WiNDR root@khalid:~# vserver status
1137017716 M * SiD3WiNDR Can not find a vserver-setup at '/etc/vservers/status/'.
1137017719 M * SiD3WiNDR maybe <vservername> is missing in the usage help? ;)
1137017764 J * eric ~eric@ebiederm.dsl.xmission.com
1137017812 M * Aiken (root@pebbles) vserver hoppy status
1137017813 M * Aiken Vserver 'hoppy' is running at context '49154'
1137017813 M * Aiken Number of processes:  2
1137017813 M * Aiken Uptime:               21:20
1137017818 M * Bertl SiD3WiNDR: yup, pls file a bug report
1137017831 M * brc Bertl: ok!
1137017833 M * Bertl welcome eric!
1137017851 M * eric hello 
1137017854 M * SiD3WiNDR Bertl: okay, will do tomorrow, am almost off to bed :)
1137017878 M * brc Bertl: that way it will be much faster than XEN
1137017886 M * eric I've just been told by email that discussions about the vserver are probably better done via irc.
1137017894 M * Bertl eric: indeed ... :)
1137017940 M * Bertl eric: so, what are your plans regarding mainline virtualization?
1137017987 M * eric The short version is I plan on attempting to suck up all of the remaining clone bits :)
1137017988 M * Bertl eric: and do you know of the 'many' approaches folks have taken and are still taking ...
1137018040 M * eric I have a clue.  I've looked at several of them and decided something better could be done.
1137018051 M * Bertl eric: hmm .. please elaboreate ...
1137018058 M * Bertl *elaborate even
1137018073 Q * Viper0482 Remote host closed the connection
1137018099 M * eric My key idea is multiple instances of global kernel namespaces. 
1137018127 M * Bertl okay, that is somewhat along the lines we do it ...
1137018143 M * eric The work on each namespace can be implemented and merged indpendently.
1137018143 M * Bertl what do you 'consider' kernel namespaces btw?
1137018169 M * Bertl ah, okay, I get the idea ... a CLONE_IPC e.g.
1137018222 M * Bertl (or CLONE_PID :)
1137018230 M * eric A little more along the lines of CLONE_NEWNS but yes.
1137018256 A * SiD3WiNDR requests CLONE_BERTL
1137018260 M * eric Currently I have been working on a proof of concept implementation to see where this takes me.
1137018268 M * Bertl okay, sounds reasonable and very compatible to Linux-VServer :)
1137018350 M * eric The key difference with my CLONE_NPSPACE ( for pids) is that  I can have multiple instances of any pid not just pid 1.
1137018372 M * eric SYSVIPC was easy.
1137018381 P * undefined 
1137018383 M * Bertl yes, that's called pid virtualization (in contrary to our isolation)
1137018391 P * stefani I'm Parting (the water)
1137018438 M * Bertl eric: and that's a requirement for snapshoting and migration, I know
1137018469 M * eric In use yes.  But I processes only have the one pid.  My process spacess are essentially disjion.
1137018506 M * eric All I have is a pid that stands in the parent process space for all of the other processes.
1137018557 M * Bertl k, so, descendants of the parent will see them, other children not
1137018576 M * eric Pretty much.
1137018591 M * eric Making proc work in that context was fun :)
1137018660 M * Bertl hmm, should be 'quite' easy if you have a 'simple' visibility check
1137018685 M * eric I am currently working on making what appears to be a second instance of the network stack.
1137018692 M * Bertl we do similar for the isolated pid spaces
1137018746 M * eric There were 2 things with proc.  If you aren't careful the  dcache will mess you up.  The other was I made added a directory on the standin pid pspace where all of proc for the child process space shows up.
1137018786 M * Bertl ah, cool, was thinking about something like that for the future (with pid virtualization)
1137018819 M * Bertl eric: okay, let's put the technical things aside for a moment, (we'll get to that automatically, I'd say)
1137018819 M * eric If you want to migrate sockets you need to take your ip addresses with you.
1137018840 M * Bertl let's talk a little about your plans regarding mainline
1137018843 M * eric Ok.  On the technical front we seem to be pretty much in agreement.
1137018883 M * Bertl yes, it sounds reasonable to cooperate and use the result for linux-vserver wherever possible
1137018900 M * eric I need to get a little farther along and then start this conversation on the kernel list.
1137018908 M * Bertl you can also 'abuse' the devel/experimental branch to test your stuff
1137018934 M * eric That is the big benefit to me in testing :)
1137018953 M * eric s/testing/cooperating/
1137018976 M * eric Plus an implementation people can agree upon is much more likely to get merged.
1137018990 M * eric So how would I abuse the dev/experimental branch.
1137018993 M * Bertl do you know of the folks at columbia doing zap and pods and this stuff?
1137019014 M * Bertl (sorry for jumping from topic to topic :)
1137019030 M * eric I have read some of their papers.  
1137019062 M * eric My impression was that it was largely an academic effort not focused on actually getting merged and being a solution.
1137019094 M * eric I seem to recall their implementation not patching the kernel but living instead in a module for maintenance reasons.
1137019097 M * Bertl well, I had a chat with them when I visited columbia university in december ...
1137019145 M * Bertl and they kind-of agreed to work on realizing their stuff with/for linux-vserver in the future ... at some point
1137019167 M * eric Cool. 
1137019178 M * Bertl so, although their stuff is 'mainly academic' it seems that some of them are interested in getting it used somewhere
1137019200 M * eric I don't have any problems cooperating with them,  I just haven't percevied a benefit in pursuing it.
1137019225 M * Bertl anyway, we should get in contact with them too, so that the snapshoting and migration part will be compatible
1137019273 J * undefined ~undefined@adsl-68-93-109-94.dsl.rcsntx.swbell.net
1137019280 M * Bertl welcome undefined!
1137019287 M * Bertl eric: okay, back to mainline ...
1137019304 M * eric Feel free.  At this point my focus is to get the namespaces merged before I worry about actually worry about how to take things in and out of the kernel.
1137019305 M * Bertl eric: so one of your aims is to get that merged by Linus, right?
1137019316 M * eric Yes.
1137019329 M * Bertl well, you chances are probably not that bad ... :)
1137019350 M * Bertl (compared to the average coder :)
1137019362 M * eric True.
1137019376 M * eric I've at least been through this process once before.
1137019398 M * daniel_hozac weren't there a bunch of IBM guys working on various virtualization thingies aimed for mainline as well?
1137019432 M * Bertl yes, notably Serge Hallyn and the folks around him ...
1137019453 M * eric I think at some point they need to be pulled into the discussion as well.
1137019461 M * Bertl the problem most approaches have is that they try very hard _not_ to change anything in the kernel
1137019482 M * eric Yes!
1137019483 M * Bertl which IMHO is the completely wrong approach
1137019505 M * Bertl because this kind of isolation/virtualization has to happen in the heart of the ekrnel
1137019515 M * eric Completely.
1137019544 M * Bertl okay, so, let's note the following:
1137019547 M * eric My large goal is to structure it so I don't need to change the logic of other points.  But I have insered and export pointer dereferences everywhere.
1137019567 M * Bertl - you want to get some virtualizations into mainline
1137019576 M * eric Yes.
1137019588 M * Bertl - I'm perfectly fine with using 'sane' virtualizations when they are in mainline :)
1137019610 M * Bertl - we are both willing to spend some time on this efford ... :)
1137019629 M * eric Yep.
1137019631 M * Bertl - we want to make the stuff reasonably 'complete'
1137019664 M * eric Correct.  
1137019677 M * Bertl - I can put the result into devel (when it passes first checks) and we will get some feedback 'how well' it works ...
1137019693 M * eric Sounds good.
1137019714 M * Bertl okay, great, did I forget something?
1137019735 M * Bertl ah, yes, when do you usually work (i.e. when are you around)
1137019735 M * eric I don't think so.
1137019767 M * Bertl I usually get up around 17:00 CET and work into the morning
1137019819 M * Bertl (CET = GMT+1)
1137019820 M * eric My schedules tends to shift depending on who I am working with and what I am working on.
1137019881 M * Bertl okay (whatever that means for us) IRC is fine for you?
1137019920 M * eric Sorry I was trying to convert my hours lately into GMT...
1137019947 M * eric Irc wors.  I frequenly just use email.  But if we start working closely enough irc is fine.
1137019953 M * eric s/wors/works/
1137019965 M * Bertl okay, Serge should be around here too, so that probably helps ...
1137019998 M * eric There is one other piece of my plan to get merged into the kernel.
1137020026 M * eric present a paper at OLS this year on the basic ideas, and how they solve the general problem. 
1137020063 M * Bertl okay, so that means?
1137020091 M * eric Well not a lot except it is another forum for the discussion.
1137020098 M * eric And it should help get the word out.
1137020123 M * eric And of course I might feed the drafs your way for sanity checking :)
1137020173 M * eric The important part is that it is a good conversation start at OLS to resolve remaining technical issues.
1137020233 M * Bertl hmm, guess OLS would be nice .. have to check if somebody would be willing to sponsor that ...
1137020321 Q * monrad Ping timeout: 480 seconds
1137020332 M * eric Have you had any conversations with the openvz guys?
1137020382 M * Bertl yes, a lot ... I'm on their IRC channel, and they are here somewhere too
1137020419 M * Bertl I guess OpenVZ would also use a mainstream feature sooner or later
1137020440 M * Bertl (but they are using older kernels right now)
1137020489 M * eric I noticed that.  The important part is to get some interface reveiw of them saying they can use the features the way they get into mainstream.
1137020514 M * eric I have already borrowed ideas from both.
1137020518 M * Bertl of course ...
1137020539 P * meandtheshell 
1137020548 M * Bertl there is also FreeVPS and alexey might be interested in this too
1137020575 J * emp ~emp@70.57.239.35
1137020583 M * Bertl welcome emp!
1137020596 M * emp Hey Bertl, thanks
1137020612 M * Bertl eric: okay, I have to leave now, SO is waiting, but I'll be back in a few hours ... will you be around then?
1137020653 M * eric Probably.  I will just leave the irc window open and check back.
1137020685 M * Bertl okay, great! so cya later ...
1137020689 M * eric later.
1137020693 N * Bertl Bertl_oO
1137020763 J * mkhl ~mkhl@200-148-41-51.dsl.telesp.net.br
1137020891 M * emp any tips on running an x11 server in a vserver?
1137020901 P * undefined 
1137020989 M * daniel_hozac like http://linux-vserver.org/Vservers+and+X ?
1137021025 M * emp thanks, i think that's what i need, I didn't see that at first glance.. thanks
1137021327 M * emp where do i stick the 'CAP_SYS_RAWIO' capability?  I'm running vserver under sarge
1137021367 M * fosco hi
1137021394 M * daniel_hozac /etc/vservers/<name>/bcapabilities
1137021395 M * fosco I see in kernel configuration that "Enable dynamic context IDs" may be removed soon
1137021403 M * fosco how can I set static xid ?
1137021428 M * daniel_hozac fosco: /etc/vservers/<name>/context
1137021434 M * fosco thanks
1137021454 J * monrad ~mikkel@213083190131.sonofon.dk
1137021462 M * daniel_hozac fosco: you will still require that option though, if you're using Enrico's tools.
1137021489 M * fosco I don't know this guy. so I think I may not ;)
1137021523 M * daniel_hozac Enrico is maintaining util-vserver.
1137021529 M * fosco gargl :)
1137021549 M * daniel_hozac (every once in a while, at least ;))
1137021592 M * fosco in fact, using static xid may help me
1137021602 M * daniel_hozac yes, you should always use static xids.
1137021632 M * fosco I have some strange things with "no such file or directory" on existing files, I think this can come from they are not allowed to the good xid
1137021667 M * daniel_hozac are you mounting that filesystem with tagxid? tagxid and dynamic xids will make thing work sporadically.
1137021675 M * fosco exactly
1137021725 M * daniel_hozac chxid from the host should solve that, after giving all the vservers static xids.
1137021750 M * fosco yep ;)
1137021774 M * fosco will util-vserver use my /etc/vservers/<name>/context ?
1137021779 M * daniel_hozac yes.
1137021804 M * fosco ok
1137022175 M * fosco are dynamic xid needed for chbind to work? :\
1137022184 M * daniel_hozac the kernel option? yes.
1137022193 M * fosco ok :)
1137022442 M * emp can i install a centos client on a debian host... do i need to start by creating a skel?
1137022524 M * fosco daniel_hozac: thanks, it is using static xid now
1137022531 M * daniel_hozac emp: that, or installing rpm/yum.
1137022544 M * fosco now, I can setxid _before_ starting the vserver \o/
1137023118 M * fosco it seems to be working really a lot better :)
1137023122 M * fosco thank you
1137023197 M * daniel_hozac you're welcome!
1137023730 M * emp daniel_hozac, could you give me an example how i would use the vserver util to create a centos 4 client? :)
1137023790 M * fosco (I personally use tarballs and scripts to customise them)
1137023805 M * daniel_hozac vserver xxx build -m yum <your options> -- -d centos42