1131927305 Q * monrad Quit: Leaving
1131927869 J * monrad ~monrad@213083190130.sonofon.dk
1131928034 Q * prae Quit: Pwet
1131928917 Q * Aiken Quit: Leaving
1131929708 Q * monrad Quit: Leaving
1131929789 M * bragon )
1131930293 M * Johnnie Has anyone here had troubles with LVM on kernel 2.6.14-vs2.0.1-rc1?
1131930363 M * bragon has anywone here have linux with ppc ?
1131930375 M * bragon Johnnie: not me
1131930621 M * Bertl bragon: yup, ppc here
1131930685 M * bragon héhé
1131930701 M * bragon Bertl: i have remake the keymap of my powerbook
1131930710 M * bragon i emulate the 2nd and 3th bouton
1131930718 M * bragon but i don't manage to copy paste
1131930720 M * bragon ...
1131930724 M * bragon any idea ?
1131930841 M * Bertl well, works on my TI powerbook
1131930852 M * Bertl F11 is paste
1131930873 M * bragon i test
1131930900 M * bragon don't work :x
1131930908 M * Bertl well, that's my remapping :)
1131930928 M * bragon i have remapping too but i don't find the copy paste ...
1131930933 M * bragon i'm nervous :)
1131930977 M * Bertl hmm, maybe use xev?
1131930993 M * bragon i try
1131931036 M * bragon no :x
1131931055 M * Bertl evtest works too
1131931121 A * bragon slaps him
1131931126 M * bragon not
1131931128 M * bragon :/
1131931295 M * Bertl well, get _some_ tool to read /dev/input/*
1131931296 M * bragon thx Bertl i will search with xev more and more
1131931386 Q * mugwump Ping timeout: 480 seconds
1131931586 J * mugwump ~samv@watts.utsl.gen.nz
1131934059 J * dddd44 dhb55@218.111.178.108
1131936057 Q * dddd44 Read error: Connection reset by peer
1131936806 J * Marchildon ~nicolas@dsl.speedline207.106.electronicbox.net
1131936816 M * Bertl welcome Marchildon!
1131936826 M * Marchildon hi Bertl
1131936834 M * Marchildon guess why I'm here :)
1131936845 M * Bertl hmm, just visiting?
1131936913 M * Marchildon nah, having problems installing a vserver for the first time :-D
1131936926 M * Bertl well, that's unusual :)
1131936945 M * Marchildon ah, I think I got why, let me try
1131936987 M * Marchildon using newvserver, --vsroot should not include the name of the vserver, right?
1131937012 M * Bertl newvserver is not part of the tools, please avoid it
1131937038 M * Marchildon isn't it the debian way?
1131937047 M * Bertl tools are: util-vserver (recent version is 0.30.209)
1131937073 M * Bertl I don't know about the debian way, but the linux-vserver way at least works :)
1131937815 M * Bertl Marchildon: hmm, did I scare you?
1131938002 M * Marchildon Bertl: maybe ;-)
1131938043 M * Bertl well, wasn't my intention :)
1131938070 M * Marchildon I'm getting a "dpkg: can't mmap package info file `/var/lib/dpkg/available`: Invalid argument", and the file is there, but empty
1131938077 M * Marchildon (using newvserver)
1131938101 M * Bertl what kernel/tools are we talking about?
1131938143 M * Marchildon anyway, I tried this first, without success: "vserver elena build -m debootstrap -- -d sid -m ftp://ftp.ca.debian.org/debian/ -- --resolve-deps"
1131938172 M * Bertl that looks better to me ... what/why/how did it fail?
1131938187 M * Bertl maybe the output of testme.sh would help?
1131938238 M * Marchildon E: no such script: ftp://ftp.ca.debian.org/debian/
1131938248 M * Marchildon testme.sh ran successfully
1131938257 M * Bertl okay, can you upload the output?
1131938264 M * Bertl (pastebin.com)
1131938355 M * Marchildon http://pastebin.com/428628
1131938364 M * Marchildon thanks for pastebin, didn't know 'bout that
1131938380 M * Bertl hmm, okay, older tools too ...
1131938398 M * Bertl there should be at least 0.30.208 tools in testing/unstable
1131938419 M * Bertl of course, the 0.30.209 tools have several bugfixes over those
1131938436 M * Marchildon ok, upgrading from stable to testing
1131938473 M * Marchildon thanks for helping, btw
1131938482 M * Bertl you're welcome!
1131938561 Q * Johnnie Ping timeout: 480 seconds
1131938717 M * Marchildon ah, 0.20.208 is in unstable, not testing
1131938835 M * Bertl well, in mainline it's already outdated :)
1131939128 J * Johnnie ~john@acs-24-154-53-217.zoominternet.net
1131939330 N * nokoya nokoyaz
1131939355 N * nokoyaz nokoya
1131939384 M * Marchildon still getting "no such script" with 0.20.208
1131939489 M * Marchildon "vserver elena build -m debootstrap -- -d sid -m ftp://ftp.ca.debian.org/debian/ -- --resolve-deps"
1131939500 M * Bertl sec
1131939501 M * Marchildon -d and -m are passed to what?
1131939523 M * Bertl first, let's try with:
1131939577 M * Bertl vserver elena build -m debootstrap --hostname elena --context 42 -- -d sarge -m ftp://ftp.ca.debian.org/debian/
1131939675 M * Marchildon it seems better...
1131939683 M * Bertl http://linux-vserver.org/alpha+util-vserver
1131939696 M * Bertl (last of the create examples)
1131939702 M * Marchildon I was here: http://linux-vserver.org/Step-by-Step+Guide+2.6
1131939771 M * Bertl yes, basically your line seems right to me
1131939798 M * Bertl but no idea what the debian packages contain ...
1131940705 M * Marchildon works!
1131940720 M * Marchildon running base-config inside the vserver...
1131940751 M * Bertl good, now I've read that sid/etc/newer will require the --resolve-deps argument
1131940780 M * Bertl but maybe the build process failed because the server has no such variant?
1131941338 Q * entroposcope Remote host closed the connection
1131943146 J * entroposcope ~entroposc@user-0c992og.cable.mindspring.com
1131943426 J * dos000 ~dos000@i216-58-60-176.cybersurf.com
1131944818 M * Bertl okay, I'm off to bed now ... have a nice whatever everyone!
1131944826 N * Bertl Bertl_zZ
1131944918 Q * lilo Quit: leaving
1131944947 J * lilo ~lilo@lilo.usercloak.oftc.net
1131945561 M * Marchildon ok, I have a lot to learn, I think - don't know how to setup networking in my new vserver :-(
1131945906 Q * lilo Quit: brb
1131945922 J * lilo ~lilo@lilo.usercloak.oftc.net
1131946820 J * Aiken ~james@tooax6-233.dialup.optusnet.com.au
1131947990 Q * dos000 Quit: Leaving
1131949608 Q * Hollow Remote host closed the connection
1131949651 J * Hollow ~hollow@home.xnull.de
1131949657 M * eyck elena, nice
1131950132 Q * Hollow Remote host closed the connection
1131952912 J * dos000 ~dos000@i216-58-60-176.cybersurf.com
1131957120 J * dlippolt ~dlippolt@cpe-70-112-77-129.austin.res.rr.com
1131957184 M * dlippolt hello all u vserver fans
1131957212 M * dlippolt todays question:  what is the current story with multicast support (both sending and receiving).
1131957232 M * dlippolt i see almost nothing on the web, and only very minimal and dated info in irc logs
1131957321 J * yungyuc ~yungyuc@220-135-53-220.HINET-IP.hinet.net
1131957565 Q * Aiken Ping timeout: 480 seconds
1131958035 J * FireEgl Atlantica@2001:5c0:84dc::
1131958851 J * shedi ~siggi@tolvudeild-198.lhi.is
1131959840 M * sizo moin
1131960531 Q * yungyuc Quit: leaving
1131960840 J * Johnsie ~john@acs-24-154-53-217.zoominternet.net
1131960841 Q * Johnnie Read error: Connection reset by peer
1131960852 P * Johnsie 
1131960878 J * Johnsie ~john@acs-24-154-53-217.zoominternet.net
1131963041 Q * virtuoso Ping timeout: 480 seconds
1131963131 J * virtuoso ~s0t0na@shisha.spb.ru
1131964650 Q * marl__ Ping timeout: 480 seconds
1131964716 Q * virtuoso Ping timeout: 480 seconds
1131964822 J * virtuoso ~s0t0na@shisha.spb.ru
1131965291 Q * MostlyHarmless Ping timeout: 480 seconds
1131966309 J * mrec ~revenger@p54B03DB3.dip0.t-ipconnect.de
1131966728 Q * mrec_ Ping timeout: 480 seconds
1131967719 J * Hollow ~hollow@82.135.28.84
1131969704 J * sebi ~sebi@Fd09f.f.strato-dslnet.de
1131969801 Q * sebi_ Ping timeout: 480 seconds
1131969818 J * ^Cist ~x@p54A5512F.dip.t-dialin.net
1131970650 Q * iprone Ping timeout: 480 seconds
1131973607 J * dddd44 dhb55@218.111.178.108
1131973747 Q * dddd44 Read error: Connection reset by peer
1131976178 Q * entroposcope Remote host closed the connection
1131977298 J * Guest185 ~Miranda@83.215.237.5
1131977322 J * entroposcope ~entroposc@user-0c992og.cable.mindspring.com
1131977838 J * iprone ~iprone@65.83.231.99
1131978947 N * Bertl_zZ Bertl
1131978952 M * Bertl morning folks!
1131978958 M * mnemoc moin Bertl 
1131978999 M * Bertl dlippolt: well, should work, if you assign the multicast IPs to the guest (and will definitely work on the host)
1131979031 Q * rocket Quit: leaving
1131979052 M * Bertl Marchildon: simply add --interface (as shown on the alpha util-vserver page) to your creation process or extend the config by hand (see Flower Page)
1131979098 M * Bertl matti: package arrived! :)
1131980031 J * shed| ~siggi@tolvudeild-198.lhi.is
1131980076 Q * iprone Ping timeout: 480 seconds
1131980210 Q * shed| Quit: 
1131980221 M * ^Cist i got a problem after some massive upgrade, maybe anyone got some ideas:
1131980295 M * ^Cist first the working environment: P4, 2GB RAM, 2.6.10-vs1.9.4, running domino within a vserver - works fine
1131980365 M * Bertl good ...
1131980388 M * ^Cist AMD Dual-Opteron, x86_64, SMP, 4GB RAM, 2.6.14 vs2.1.0rc5, running the same domino vserver (32bit) starts and runs fine for a while but with many users domino stops working
1131980482 M * Bertl hmm, stops working? maybe it's tired? :) seriously, can you nexplain the stops working part?
1131980495 M * Bertl *explain
1131980521 M * Bertl also, have you tried the stable branch (vs2.0.x) yet?
1131980536 M * ^Cist which means as far as i could debug, that the Recv-Q of the clientconnections don't get read by the domino servers anymore
1131980575 M * ^Cist domino is around 200-300 processes here, so it seems that it partly dies - some of these processes keep working
1131980674 M * ^Cist nope, havent tried another patch yet because i wanted to ask here/you first since this machine is in production so i want to avoid downtimes
1131980702 M * ^Cist what are (in short) the differences between 2.0 and 2.1? 
1131980744 M * Bertl 2.0.x is the stable branch, it is designed to just work, the 2.1.x is the development branch, many new features, but not everything is expected to work
1131980770 M * ^Cist this, i know :) - i meant more technically
1131980828 M * Bertl hmm,more technically speaking, it's roughly this:
1131980830 M * Bertl http://vserver.13thfloor.at/Experimental/delta-vs2.0.1-vs2.1.0.diff.bz2
1131980901 M * dos000 lol
1131980909 M * ^Cist ok, so it's either rtfs or one is devel and the other is supposed to be stable
1131980926 M * Bertl (BME, CoW, KT, some accounting, different locking)
1131980949 M * Bertl there will be a feature list pretty soon, didn't get to write it yet
1131980987 M * Bertl ^Cist: if you scan the IRC logs, you can probably get a nice description of each feature
1131980994 M * ^Cist ok, the locking part is something that would have made me try anyway :)
1131981031 M * Bertl I'd advise to at least update to the latest devel release
1131981437 Q * dos000 Quit: Leaving
1131981736 M * derjohn how do I set a cpu limit for a server? run-time? new-style config? I compiled "hard cpu limit" in ... is the opposite a "soft one" or "none"?
1131981807 M * derjohn EHLO by the way :)
1131981860 M * Bertl yes, yes, yes and yes :)
1131981879 M * derjohn foobar!
1131981923 M * Bertl you can set it at runtime (as most options), you can also configure it in the config
1131981947 M * Bertl and the alternative to hard-cpu (choosable via flag) is priority scheduler or none (normal scheduler)
1131981948 M * derjohn to be precise in my question: how? vdlimit?
1131981968 M * Bertl that's disk limits :)
1131981970 M * derjohn Bertl, on WTH you demonstrated the hard cpu limiter?
1131981977 M * Bertl vscked ...
1131981981 M * Bertl *vsched
1131981982 A * derjohn slaps himself
1131982002 M * Bertl yes, on WTH we demonstrated the hard scheduler too
1131982004 M * derjohn where can I get that "cpuhog" util you used?
1131982139 M * Bertl http://vserver.13thfloor.at/Experimental/TOOLS/cpuhog.c
1131982194 M * derjohn ah, fine, thx. didn't notice tat it is your stuff!
1131982201 M * derjohn btw: vsched [--xid <xid>] ....
1131982229 M * derjohn it is very confusing, to talk about ctx, context and xid.
1131982246 M * matti Bertl: All OK?
1131982260 M * derjohn i expected vsched to live in te world of context's, not xids ...
1131982263 M * Bertl matti: didn't test yet, but looks like ... TX!
1131982283 M * Bertl derjohn: well, it's very simple actually:
1131982285 Q * Pazzo Quit: ...
1131982298 M * Bertl - ctx is legacy talk, don't use it
1131982322 M * Bertl - context is something which might be a process or network isolation
1131982335 M * Bertl - xid is the context id for process contexts
1131982361 M * Bertl (the x in xid is from conteXt :)
1131982378 M * derjohn Bertl, I am used to live wit it. XID Xtended-context ID. :)
1131982401 M * Bertl well, nothing extended there, though ...
1131982420 M * derjohn in case of file-systems it is eXtented....
1131982433 M * Bertl like in xfs? :)
1131982433 M * derjohn tat wy I thought ... well ... I should stop thinking too much.
1131982435 M * matti Bertl: OK.
1131982445 M * Bertl matti: will test later today ...
1131982450 M * matti Bertl: Aye.
1131982460 J * Pazzo ~Pazzo@host130-250.pool8172.interbusiness.it
1131982465 M * matti Bertl: I hope, that HDD will not be broken, or someting :)
1131982469 M * derjohn [--tokens-max <tokens>] ... er, how do I figure out how may tokens there are on my kernel. jiffies?
1131982530 M * Bertl http://linux-vserver.org/Scheduler+Parameters
1131982546 M * derjohn Bertl, do you run a bot for these answers on IRC ? :)
1131982589 M * Bertl although I was suspected several times to actually _be_ a bot ... no, it's quite simple, 3 open browser windows are more than sufficient
1131983143 M * derjohn Bertl, the schedular page os not very self-explaining. would that be right: vsched --xid 125 --fill-rate 7 --interval 32 --tokens 500 --tokens-min 200 --tokens-max 1000
1131983153 M * derjohn for a "20% of 1 CPU case" ?
1131983163 M * derjohn what is "dummy" for?
1131983229 M * Bertl an unused line
1131983238 M * Bertl yes, should be roughly 21%
1131983247 M * Bertl (in the saturated case)
1131983271 M * derjohn in the conf file? ok. I thought it was the name of some scheduler, like "dummyjiffies" or so
1131983296 M * derjohn with ard cpu limit I only get 21-hard-% ?
1131983306 M * derjohn even if the rest is unused?
1131983404 M * Bertl yes, that's the idea
1131984614 Q * Guest185 Read error: Connection reset by peer
1131985835 Q * shedi Quit: Leaving
1131986446 M * ^Cist ok, 2.6.14.2-vs2.0.1rc2 now running... i'll let you know if it works or fails again
1131986786 M * Bertl excellent, thanks!
1131986819 M * Bertl off for dinner now ... back shortly
1131986823 N * Bertl Bertl_oO
1131986842 M * ^Cist nox, it's me who says thanks :)
1131986852 M * ^Cist s/nox/no
1131986852 M * derjohn ^Cist, be a brave man, try 2.1rc7 !
1131988393 Q * cereal Read error: Connection reset by peer
1131988551 Q * baggins Ping timeout: 480 seconds
1131988836 J * iprone ~iprone@65.83.231.99
1131988853 J * cereal koepi@217.20.124.153
1131989185 J * Viper0482 ~Viper0482@p549768BF.dip.t-dialin.net
1131989426 J * MostlyHarmless ~mh@melbourne.mostly-harmless.ca
1131989597 J * PhiberOptics ExUser@adsl-68-89-42-36.dsl.hstntx.swbell.net
1131989798 N * Bertl_oO Bertl
1131989805 M * Bertl back now ...
1131989816 M * Bertl hmm, welcome PhiberOptics!
1131990173 M * Pazzo Hi Bertl!#
1131990242 M * Bertl hey Pazzo! LTnS!
1131990249 Q * sannes Quit: BitchX-1.1-final -- just do it.
1131990290 M * Pazzo yeah - little bit too long :-)
1131990300 M * Pazzo how are you? everything fine?
1131990349 M * Pazzo probably I missed lots of things...
1131990362 M * Bertl everything :)
1131990409 M * Pazzo hehe - at least the flower page still looks the same ;-)
1131990425 M * Bertl lol
1131990433 M * matti LOL
1131990485 M * Pazzo where can I read up everything about 2.1.0-rc7?
1131990492 M * Pazzo hey, already rc7?
1131990506 M * Pazzo wow... 
1131990522 M * Bertl here: http://vserver.13thfloor.at/Experimental/patch-2.6.14.2-vs2.1.0-rc7.diff
1131990544 M * Bertl (it contains all the dirty little details :)
1131990555 M * Pazzo hehe
1131990565 M * Pazzo maybe a changelog?
1131990576 M * Pazzo or still RTFIL?
1131990580 M * Bertl no, I guess I have to add something, maybe the feature matrix?
1131990583 M * Pazzo IL=> IRC Logs
1131990607 M * Bertl yes, didn't bother to do a changelog yet (thought I wait for release)
1131990623 M * Pazzo feature matrix?
1131990739 M * Pazzo is NGN part of 2.1.x?
1131991020 M * Bertl well, it is slowly getting part of it ...
1131991033 M * Bertl (means: we are currently redesigning ngnet)
1131991112 M * Pazzo redesigning?
1131991179 M * Bertl yeah, well, making it even more flexible than the first version :)
1131991208 M * Pazzo is there a way for me to find out details / get a big picture of where ngnet is moving (avoiding asking you too many already-ansered questions but without reading the source code) ?  :-)
1131991321 M * Bertl hmm, I guess not, but you can try out some things if you like :)
1131991361 M * Pazzo are you still using iptables to mark packages?
1131991370 M * Bertl nope
1131991407 M * Pazzo ok, so I probably have absolutely no idea what NGNET looks like today
1131991418 M * Bertl hmm, yep, probably :)
1131991438 J * menomc ~amery@200.75.27.74
1131991455 Q * mnemoc Read error: Connection reset by peer
1131991455 N * menomc mnemoc
1131991517 M * Pazzo how are packets handled now? I mean: are there virtual interfaces (allowing per vserver ipbables?)? do packets behave like in the current stable implementation or totally different? what about routing tables?
1131991542 M * Pazzo is http://linux-vserver.org/NGNET-Testing-HOWTO up2date?
1131991555 M * Bertl first, interfaces are not related in any way to iptables or routing tables
1131991586 M * Pazzo yeah, I know that - you personally teached me about this :-)
1131991623 M * Bertl okay, so there are virtual interfaces, and it will look more like ppp or tun/tap
1131991632 M * Pazzo "teached me about this" - sounds like really bad English - sorry
1131991656 M * Bertl of course, the final version will include virtualized tables (routing/netfilter) too
1131991683 M * Pazzo cool, sounds great! If I remember it right you have always been against tun/tap-like interfaces?
1131991703 M * Bertl well, because of the overhead ... but I figured two things since
1131991724 M * Bertl a) folks do not care about overhead when they are looking for virtual devices
1131991734 M * Pazzo right :)
1131991739 M * Bertl b) I figured a method to keep the overhead pretty low
1131991773 M * Pazzo without copying packets to userspace?
1131991776 M * Bertl as I said, if you like you can do a simple test (basically just a proof of concept)
1131991785 M * Bertl yes, no userspace involved
1131991806 A * MooingLemur pretends to moo, for no reason.
1131991839 M * Bertl MooingLemur: shouldn't that be '..., for no apparent reason.' ?
1131991846 M * Pazzo Bertl: I would like to - but NGNET-Testing-HOWTO seems a little bit outdated
1131991849 M * MooingLemur probably :)
1131991872 M * Bertl Pazzo: well, documentation is outdated the day you write it, that's why folks come here ...
1131991873 M * Pazzo apt-get moo
1131991915 M * MooingLemur emerge moo
1131991941 M * Pazzo ok Bertl, so I'll not let you in peace: where do I start?
1131992021 M * Pazzo and: I know that there is no timeline - but what do you think when a) 2.1 will be released as "stable" and b) will NGNGNET then be there?
1131992078 M * Pazzo do I need patch-2.6.14.2-vs2.1.0-rc7.diff and delta-2.6.14-vs2.1.0-ngn0.02.diff ?
1131992083 M * Pazzo and what about userspace?
1131992085 Q * cryo Remote host closed the connection
1131992085 M * Bertl I'd say 2.1 will be released within the next month ..
1131992166 M * Pazzo hmm... "vdevtag" (found on Experimental/NGNET/dualhome.info) looks pretty new
1131992208 M * Pazzo delta-ngn0.03.diff or delta-2.6.14-vs2.1.0-ngn0.02.diff?
1131992275 M * Bertl just a moment, *phone call*
1131992516 M * Pazzo ok, I downloaded (and compiled) vnet.c and vdevtag-0.02.tar.bz2
1131992543 M * Pazzo then downloaded delta-ngn0.03.diff and startet downloading 2.6.14.2
1131992730 M * PhiberOptics help: i need to chmod /var/www/html/ so regular users can add and change files via ftp
1131992750 M * PhiberOptics what is the command?
1131992781 M * Bertl chmod?
1131992795 M * Bertl Pazzo: that's an excellent choice ...
1131992801 M * Pazzo thnx :)
1131992829 M * Pazzo Bertl: any special utils version required?
1131992900 M * Bertl nope, 0.30.209
1131992910 M * Pazzo perfect
1131992915 M * Bertl you need to get the vdevtag 0.02 and the vnet.c
1131992920 M * Bertl and compile both
1131992930 M * Bertl also make sure that you have local access to the amchine
1131992936 M * Bertl (or serial console)
1131992965 Q * PhiberOptics Quit: « Ë×Çü®§îöñ » Info~[v9.5]~ Released~[October 27, 2003]~
1131992967 M * Pazzo so I'll build my 2.6.14.2-vs2.1-rc7-ngnet02 debian package, install it, reboot...
1131992977 M * Pazzo ...vdevtag and vnet are compiled and ready
1131993046 M * Bertl okay, perfect ...
1131993066 J * PhiberOptics ExUser@adsl-68-89-42-36.dsl.hstntx.swbell.net
1131993075 M * Bertl wb PhiberOptics!
1131993105 M * Pazzo great - I have somehow broken my dependencies with zaptel-modules **grrr**
1131993166 M * Pazzo ok, fixed
1131993263 M * Bertl let me know when you are booted and have a console @ hand
1131993265 M * Pazzo patches applied, no errors / notices
1131993324 M * Pazzo what about kernel config? anything special?
1131993399 M * Bertl ah, yes, you need to enable NGNET
1131993435 M * Pazzo (I've chosen PREEMPT_NONE)
1131993474 M * Pazzo HZ_1000?
1131993488 M * Bertl won't hurt ...
1131993509 M * Bertl derjohn: hmm, wouldn't the FAQ be better linked with the other FAQs?
1131993773 J * dddd44 dhb55@218.111.178.108
1131993828 J * stefani ~stefani@superquan.apl.washington.edu
1131993995 M * Pazzo cool - is VSERVER_COWBL ready to be seriously used?
1131993997 J * cryo ~say@212.86.233.146
1131994067 M * Pazzo hmmm.. INOXID_RUNTIME is new - should I stay with INOXID_UGID24? I'm currently not using disk quotes / quotas inside vservers - but if there has been some improvement I would really like to do so
1131994102 M * Pazzo and what about XID_PROPAGATE?
1131994120 M * Pazzo Bertl: seems that I missed a lot of new things :-)
1131994194 M * Pazzo Disable Legacy Networking Kernel API => YES
1131994235 Q * Viper0482 Remote host closed the connection
1131994241 M * Pazzo should I enable VSERVER_HARDCPU_IDLE?
1131994260 M * Bertl no neccesary
1131994282 M * Pazzo not for NGNET - but generally?
1131994307 M * Pazzo I  would also like to test the hard cpu scheduler as it seems to be really stable now, right?
1131994329 M * Pazzo so if I would like to do so - should I also enable VSERVER_HARDCPU_IDLE?
1131994574 M * Pazzo and could you be so kind to also give me a short hint regarding the other questions?
1131994707 M * Bertl yes, but you won't be able to test much with the ngnet patches applied
1131994713 M * Pazzo ?
1131994741 M * Bertl you probably want to test the hard cpu _without_ the ngnet patches
1131994758 M * Bertl (otherwise your guests will not have network connectivity)
1131994780 M * Pazzo no network connectivity with ngnet?
1131994789 M * Bertl yeah, kind of
1131994800 M * Pazzo great :-)
1131995091 M * Pazzo Bertl is once again abusing me for animal experiments ;-)
1131995115 M * Pazzo kernel package is being built - I'll make a coffee-break
1131995173 M * Pazzo (building will probably take some time as it is running on a slow host)
1131995218 Q * dddd44 Read error: Connection reset by peer
1131995269 M * Pazzo BBL
1131995512 A * Bertl would never do animal experiments .. that's what humans are foe :)
1131995564 M * Bertl *for
1131996146 M * michal_ lol
1131996399 M * matti well... I've third eye ;]
1131996400 M * matti ;p
1131996538 M * michal_ wow
1131996541 Q * iprone Ping timeout: 480 seconds
1131996545 A * michal_ have done skype talk with sister
1131996547 M * michal_ amazing things ;p
1131996586 M * michal_ i need some headset probably - mic and speakers are build into the powerbook and it is a bit tricky to set it in this way it won't interfere
1131996607 M * Bertl well, there are good open source solutions too ...
1131996654 M * michal_ tell me about one that is also supported under linux @ ppc
1131996661 M * michal_ and windows
1131996665 M * michal_ and mac os :)
1131996665 M * mnemoc any SIP client should work :)
1131996680 M * Bertl open h323 too ...
1131996710 M * mnemoc there is no excuse to use skype
1131996716 M * michal_ lol
1131996727 M * michal_ i feel quite good about it
1131996737 M * michal_ and have already friends using it so you know...
1131996747 M * matti mnemoc: I don't have a microphone ;]
1131996756 M * michal_ the pitty is that i cannot talk when working under linux
1131996772 M * Bertl matti: hmm, thought you went to bed :) well, you can use festival :)
1131996785 M * daniel_hozac haha
1131996814 M * matti Bertl: Yeah... I am in the bed... Laptops are evil.
1131996820 M * michal_ lol
1131996831 M * michal_ matti your cat is somewhere there ?
1131996836 M * michal_ one of them actualy
1131996837 M * matti michal_: Yep.
1131996837 M * Bertl michal_: KPhone should work on all paltforms :)
1131996850 M * michal_ Bertl: will take a look
1131996868 M * matti michal_: Want something from my cat? :)
1131996879 M * michal_ sure
1131996894 M * Bertl you might need to install a larger part of cygwin on windows, but hey, windows users are used to larger libraries :)
1131996911 M * michal_ sister has almost empty disk so you know ;]
1131996921 M * matti LOL
1131996929 M * michal_ stroke that cat from me matti ;p
1131996931 M * mnemoc google talk 
1131996970 M * michal_ hm. why do you hate skype (except it is not free ?)
1131997015 M * Bertl don't hate it, you got that wrong, just I do not want to use/install software I do not know anything about
1131997035 M * michal_ quite valid point
1131997045 M * Bertl migh easily send my private files to whoever ...
1131997071 M * michal_ i have once did installed opera under rsbac machine and have enabled auditing on this application
1131997078 M * michal_ it was not behaving friendly ;]
1131997095 M * michal_ rotfl ! overkill of a year ;p
1131997097 M * michal_ look here
1131997103 M * michal_ http://store.apple.com/1-800-MY-APPLE/WebObjects/AppleStore.woa/70607/wo/Zvx7eobiHWLZ2F9rFUWwSKaCNg3/7.SLID?mco=E02A4992&nplm=TF410LL%2FA
1131997116 M * michal_ (uh, sorry, should have been made it shorter)
1131997191 M * matti You're kiddin; me...
1131997193 M * matti ROTFL
1131997581 J * kjo ~krischan@csd-ip-139-7-62-170.vodafone-net.de
1131997583 M * FaUl someone how knows how the linux-memory-menagement works?
1131997588 Q * kjo Quit: 
1131997607 M * michal_ the god himself ;]
1131997638 M * FaUl i'd like to know how I have to use the brk()-call
1131997664 M * matti FaUl: There's a good documentation on that.
1131997683 M * FaUl and wether the have to brk() the stack-memory 
1131997688 M * FaUl matti: fine, where?
1131997699 M * FaUl s/the/I/
1131997741 M * matti FaUl: Heh, you not even ask the Google...
1131997746 M * matti FaUl: http://www.csn.ul.ie/~mel/projects/vm/
1131997790 M * Bertl FaUl: sys_brk() just changes the VM assignment for a process
1131997906 M * FaUl Bertl: I guess I'll read the link matti gave to me
1131997925 M * Bertl probably a good idea :)
1131997966 M * michal_ hm, i will print it out
1131997995 M * matti michal_: Some parts of Linux VM are described in the book, that I gave you :)
1131998007 M * matti michal_: Did you use this book even? ;]
1131998019 M * michal_ i am using it all the time !
1131998026 M * michal_ man, i would not exist without it
1131998047 M * matti Heh, nice.
1131998057 A * Bertl .o( hmm, support for a broken table? :)
1131998083 M * michal_ this paper is more detailed, but yes, i have been reading about mm there. (i had no idea what slabs were before ;p)
1131998214 J * schak schak@dslb-082-083-043-091.pools.arcor-ip.net
1131998225 M * Bertl welcome schak!
1131998260 M * mnemoc can i give a capability to a running context?
1131998279 M * schak hi Bertl, hi @all
1131998281 M * Bertl ccaps immediately, bcaps yes, but will require a fork
1131998346 M * mnemoc a fork?
1131998364 M * Bertl well, the bcaps are stored in the task struct
1131998374 M * Bertl the context only has the 'upper' limit
1131998413 M * Bertl so ... you will need to fork/exec an sud binary to get them
1131998425 M * mnemoc new processes
1131998949 J * mef ~mef@targe.CS.Princeton.EDU
1131998958 M * mef hello
1131998965 M * Bertl welcome mef!
1131998968 M * mef whoops... let me check on the console.
1131999026 M * mnemoc Bert: another dumb question, if make a loop device on the guest with the same minor than one outside or on another guest... will they be the same device if i mount something there?
1131999040 M * Bertl yep
1131999079 M * mnemoc uhm, ok.
1131999705 J * monrad ~monrad@213083190130.sonofon.dk
1131999711 M * Bertl welcome monrad!
1131999728 M * monrad evening
1132000624 M * Bertl hmm, does anybody know if qmake runs without qt?
1132000671 M * Bertl (and btw, where I would get it without the qt blob)
1132000758 M * mnemoc http://websvn.kde.org/trunk/qt-copy/qmake/?rev=480412
1132000807 M * Bertl k, tx
1132000834 M * mnemoc how should i copy a loop to have it usable?
1132000841 J * Aiken ~james@tooax6-207.dialup.optusnet.com.au
1132000848 M * Bertl morning Aiken!
1132000854 M * Bertl mnemoc: usable for what?
1132000876 M * mnemoc Bertl: losetup-able inside the guest :(
1132000886 M * michal_ cp -a will do the job
1132000887 M * mnemoc mknod doesn't seem to work for me :(
1132000892 M * Bertl cp -va  for example
1132000899 M * mnemoc i tried....
1132000901 M * mnemoc .oO
1132000908 M * Bertl but mknod will work too, you ahve to do it on the host, though
1132000939 M * michal_ vserver security ^^ (you do not want anybody to use it inside a guest, really ;)
1132000946 M * Aiken hi
1132000969 M * mnemoc my build server, which is a guest, needs a loop to populate initrd :p
1132000977 M * Bertl well, with proper secure mount and that, it should  not be that problematic
1132001007 M * michal_ i mean creating devices in general, not this particular case ;]
1132001068 M * Aiken looks like I get to compile 3 new kernels, just noticed 2.1.0-rc7
1132001220 M * Bertl hmm, just minor changes, but one interesting fix
1132002042 J * yarihm ~yarihm@80-218-5-17.dclient.hispeed.ch
1132002767 M * Aiken what is the fix?
1132002824 M * Bertl http://vserver.13thfloor.at/Experimental/FOR-2.0.1/delta-indev-fix01.diff
1132002836 M * Bertl (a bug introduced by the 2.6.14 port)
1132002926 J * MacTen macten@81.25.47.194
1132002942 M * Bertl welcome MacTen!
1132002955 M * Aiken it took until 2.6.14.2 to fix the alpha bug that was introduced into the kernel going from 2.6.14-rc3 to release
1132002989 M * MacTen hello. can any one can tell me desrciption of some "bug-feature" of vserver
1132003016 M * Bertl MacTen: probably :)
1132003020 M * MacTen i have done echo "^19" tp flags file
1132003028 M * MacTen then restart vserver
1132003061 M * MacTen so, when i type "top" at main (physical) server it must show load average for server
1132003089 M * MacTen but when i type "top" inside vserver it will show load average for this vserver, not for physical server..
1132003096 M * MacTen am i right?
1132003150 M * Bertl yup, but 'virt_load' would do too ...
1132003167 M * MacTen but why i see sometime
1132003173 M * MacTen this one.. for example
1132003174 M * Bertl (or is supposed to work, insted of ^19)
1132003187 M * MacTen top at main server shows 1,01
1132003201 M * MacTen and top inside of one of vservers show 1,15
1132003229 M * MacTen how laod average for vserver can be more than load average of physical server?
1132003274 M * MacTen what is "virt_load" ?
1132003351 M * Bertl virt_load is the human readable name for ^19
1132003373 M * Bertl and the host context only calculates the load average of host processes IIRC
1132003418 M * Bertl (not 100% sure on that)
1132003516 M * MacTen so, what is the best way to monitor load average for main server and for each vserver?
1132003551 M * Bertl you get the load of the guests in the /proc/virtual too
1132003602 M * Bertl yes, just checked, the load is evaluated from the nr_running()
1132003719 M * MacTen Bertl, shal i use cvirt at /proc/virtual
1132003804 M * Bertl yup, that should show the load inside
1132003870 M * MacTen i will try it. so in this case in can't happen' when LA of some vserver (or sum of LA os all vservers) can be more then LA of main server?
1132003884 J * shedi ~siggi@inferno.lhi.is
1132004034 M * Bertl MacTen: I think this is really a bug-feature ... 
1132004085 M * MacTen Bertl, so what do you think - should i rewrite my script to using "cvirt" from /proc/virtual or still using /proc/loadavg for each vserver?
1132004101 M * MacTen what will be more exact?
1132004149 M * Bertl you are definitely better off when using the /proc/virtual
1132004158 M * Bertl as you do not have to enter the guest
1132004167 M * Bertl but the results will be somewhat similar
1132004168 M * MacTen ok. thanks
1132004198 M * Bertl I'm now doing some checks regarding the load average
1132004219 M * MacTen i wrote my own monitoring engine for my server
1132004230 M * MacTen and saw this bug-feature :)
1132004295 M * Bertl yes, I think (at first glance) the guest is updated more often than the host
1132004312 M * Bertl MacTen: what HZ value do you use? and what kernel/patches?
1132004368 Q * MacTen Quit: Peace and Protection 4.22
1132004781 J * MacTen macten@81.25.47.194
1132004786 M * MacTen Bertl, hey
1132004793 M * MacTen i modified my script
1132004799 M * MacTen anyway i got a bug :)
1132004800 M * MacTen avg_main=0.54|#leo#avg=0.56
1132004833 M * MacTen main server LA was 0.54 and at this time LA of one VPS was 0.56
1132004835 M * Bertl yes, I see similar here, but, it looks like the guest average is just over a smaller amount of time
1132004911 M * MacTen one thing is better tha i don't need to enter the guest
1132004934 M * Bertl so, what HZ value do you use?
1132004946 M * MacTen Bertl, how can i look?
1132004953 M * MacTen i am not a profi....
1132004964 M * MacTen :)
1132004970 M * Bertl okay, check the .config in /boot or if you have the config.gz in /proc?
1132004994 M * Bertl or if you bilt the kernel yourself, in the kernel source tree (.config)
1132005005 M * MacTen no config.gz in /proc
1132005032 J * iprone ~iprone@lawn-199-77-214-191.lawn.gatech.edu
1132005044 M * Bertl welcome iprone!
1132005096 M * Bertl MacTen: because what I see here, the intervals differ, not the actual results
1132005121 M * Bertl started 4 cpuhog's in a test guest ...
1132005133 M * MacTen where i should look for .config?
1132005145 M * MacTen /lib/modules/2.6.12-1.1376_FC3/build/.config
1132005145 M * MacTen /lib/modules/2.6.9-1.667/build/.config
1132005161 M * Bertl uname -a contains one of those names?
1132005172 M * Bertl 2.6.12-* or 2.6.9-*
1132005183 M * MacTen Linux 2.6.11-vs1.9.5-rc1 #4 SMP
1132005186 M * MacTen no
1132005225 M * Bertl so you probably want to: a) update to 2.0 soon, and b) look for the kernel config of your currently running kernel
1132005243 M * Bertl but I guess it uses 1000Hz, as the HZ changed happened after 2.6.11 IIRC
1132005254 M * MacTen /usr/src/linux-2.6.11/.config
1132005263 M * Bertl looks good
1132005272 Q * yarihm Quit: Leaving
1132005287 M * MacTen CONFIG_KERNEL_HZ=100
1132005289 M * MacTen is it ?
1132005309 M * Bertl yep, that's it, so 100Hz
1132005357 M * MacTen is it bad?
1132005373 M * MacTen 100Mhz is not much as i think..?
1132005393 M * Bertl nope, that's quite fine
1132005407 M * MacTen i have celeron 2,4Ghz server
1132005408 M * Bertl 100 Hz, no mega there ;)
1132005598 M * MacTen ough, yeah
1132005611 M * MacTen not Mhz, just Hz :)
1132005629 M * Bertl is the system under heavy load atm?
1132005638 M * MacTen no
1132005647 M * Bertl could you do a few tests for me?
1132005658 M * MacTen which one?
1132005691 M * Bertl I'd like you to start a process which consumes all CPU in one guest (preferable an otherwise empty silent one)
1132005711 M * MacTen and?
1132005722 M * Bertl and check precisely _how_ long it takes until the guest load reaches 1.0
1132005748 M * MacTen very fast
1132005756 M * Bertl and _how_ long it takes for the host load to climb up 1 unit
1132005772 M * MacTen acerage LA is about .90 right now
1132005774 M * Bertl it is supposed to be 60 seconds wall time
1132005792 M * Bertl I can provide such a process to you (cpuhog)
1132005835 M * Bertl http://vserver.13thfloor.at/Experimental/TOOLS/cpuhog.c
1132005846 Q * MacTen Read error: Connection reset by peer
1132006034 M * mef bertl: need to chat with you immediately, as we need to finalize ticketing.  More importantly, can you get morrigan too?
1132006052 M * mef bertl: the ckrm machine is back up.
1132006087 M * Bertl mef: let me check ...
1132006109 Q * tchan Quit: WeeChat 0.1.7-cvs
1132006230 Q * sladen Ping timeout: 480 seconds
1132006264 J * tchan ~tchan@c-67-174-18-204.hsd1.il.comcast.net
1132006485 J * sladen paul@starsky.19inch.net
1132006877 Q * wolog Remote host closed the connection
1132006948 M * Bertl mef: regarding ckrm, what was it?
1132007314 J * arok ~arok@p54944349.dip.t-dialin.net
1132007337 M * arok hi there ^^
1132007480 M * mef bertl: need an answer in the other channel
1132007515 M * arok i need help with my vserver... i have to build tunnels (tun) from inside the vservers... i already added NET_ADMIN and NET_RAW and created /dev/net/tun but it does not work. the tun interfaces appear on the host but not on the vserver which build it. anyone can help me?
1132007521 M * Bertl welcome arok!
1132007569 M * Bertl arok: hmm sounds interesting ...
1132007609 M * arok it's not a usual tunnel... it is a so called gtp-tunnel.. but should make no difference... i think
1132007624 M * sizo re
1132008192 M * mef bertl: sorry, I did not write down the problem.
1132008197 M * arok the main problem is that i can't configure the tun interface in /etc/vservers/foobar/interfaces/ because it's not established when i start the vserver. the vserver itself has to establish the tunnel. the vserver can establish the tunnel but can't configure it because it is not listed with "ip addr" on the vserver. but it is on the host...
1132008212 M * mef bertl: there was a panic in mount or something.  Doing some munmap.
1132008220 M * Bertl mef: okay, how to proceed? because the next reboot will probably hang there again?
1132008228 M * mef bertl: sorry that I don't recall the details.
1132008241 M * mef bertl: I never got around to fixing the remote console + reboot support for you.
1132008253 M * mef bertl: unfortunately, I cannot do that tonight, as I am already late again for dinner at home.
1132008266 M * Bertl np, I'll wait :)
1132008270 M * mef Will take care of it tomorrow and then you can crash and reboot the machine to your hearts content.
1132008278 M * mef ok
1132008280 M * mef have to go
1132008280 M * mef cheers
1132008288 M * Bertl excellent! tx! cya!
1132008398 M * sizo n8
1132008415 M * Bertl night sizo!
1132008417 M * sizo bye Bertl 
1132008452 M * Bertl arok: I doubt that you will be able to establish this kind of tunnel easily
1132008472 M * Bertl arok: btw, what kernel/patch version are we talking about?
1132008558 M * arok if it would be easy i wouldn't have worked on it since 2weeks (but okay i am a vserver-noob) ;)
1132008575 M * arok uh... 2.6-?
1132008637 M * Bertl I suspect one of the main issues you encounter is, that the guest is bound to certain IPs, while the tunnel uses 'other' IPS
1132008667 M * Bertl so probably adding the tunnel IPs to the guest config, will make the interfaces appear :)
1132008695 M * arok where should i add them?
1132008715 M * Bertl to the guest config, e.g. /etc/vservers/<guest>/interfaces/
1132008736 M * arok but i had to make a interface for all possible ips, right?
1132008751 M * Bertl well, no, just for the tunnel IPs :)
1132008792 M * arok jep. that's a class a network
1132008802 M * arok at least a part of it
1132008817 M * Bertl for a test, you could narrow it down to a single one
1132008858 M * arok but i think it won't work anyway. i already picked up one ip and defined the interface in the directory
1132008863 M * arok but it didn't show up
1132008870 M * Bertl OTOH, as you already eliminated all security (with adding CAP_NET_RAW and more important CAP_NET_ADMIN) why not simply avoid the network restrictions at all?
1132008872 M * arok and the vserver complained that the interface doesn't exist
1132008901 Q * iprone Ping timeout: 480 seconds
1132008902 M * arok i don't need restrictions, there won't be any users on the vserver ^^
1132008913 M * Bertl yes, that's what I meant
1132008936 M * Bertl so why not use chcontext, and chroot/vnamespace but not chbind?
1132008969 M * arok sorry, i don't understand... (vserver-noob).
1132008973 M * ^Cist imho it doesn't make much sense to implement firewalls/gateways/routers and thelike within a vserver... it'd be more like playing for the fun of playing with vserver
1132008980 M * arok i already added net_raw and net_admin to bcapabilities
1132009036 M * Bertl what tool version do you use?
1132009042 M * arok ^Cist: i don't try to implement these things ;)
1132009056 M * arok the most recent vserver-utils i think
1132009067 M * Bertl so 0.30.209 then ...
1132009101 M * Bertl well, change the /etc/vservers/<name>/interfaces dir to the following:
1132009102 M * arok i don't remember the version. but the build date was about 28th october 2005 ^^
1132009127 M * Bertl make it contain a single dir called '0' with the following files
1132009134 M * ^Cist arok, what else is this machine / vserver supposed to do then? :)
1132009142 M * arok (i don't have access to my server here so i can only tell what i remember :( )
1132009153 M * arok *listening*
1132009184 M * Bertl ip - '0.0.0.0'
1132009204 M * Bertl prefix - '0'
1132009213 M * Bertl nodev
1132009233 M * arok ^Cist: that's something really complicated. it will be a complex sgsn-emulator sometime for testing ggsn's ^^
1132009238 M * arok okay...
1132009265 M * arok i already tried this ip but with tun1 in dev
1132009274 M * arok you mean no dev file at all or "nodev" in dev?
1132009292 M * Bertl then add to the cflags file, add ~hide_netif
1132009315 M * Bertl the file is called 'nodev' and you just need to touch it
1132009326 M * Bertl (see the Flower Page for details)
1132009329 M * arok okay.
1132009362 M * arok (the flower page would be a lot better without the flowers and colors^^)
1132009379 M * Bertl well, the choice (of stylesheet) is yours :)
1132009401 M * Pazzo re
1132009413 M * arok okay, is that everything? thanks a lot! i'll test it tomorrow :)
1132009425 M * Bertl yep, it should do the trick IMHO
1132009429 M * arok there are various style-sheets? oh .. i haven't noticed that ^^
1132009516 M * Bertl Pazzo: so how's the kernel?
1132009529 M * Pazzo its ready to run
1132009541 M * Bertl excellent! so, wanna try?
1132009570 M * Pazzo kernel-headers, -source, -image and -docs - everything build and packaged in cute .debs :-)
1132009573 M * Pazzo let's go
1132009583 M * Pazzo do I need a local console?
1132009600 M * Pazzo or will it reboot normally?
1132009605 M * arok oh... i have a second question. it's not such important yet but it would be nice to be solved ^^ if i type vserver foobar stop the vserver takes down the network interfaces after printing "rebooting..." ^^ maybe there is a trick to avoid this? ^^ 
1132009607 M * Bertl I would really advise it, but it might work a little over entwork too :)
1132009622 M * Pazzo hehe - let's give it a try
1132009645 M * Bertl arok: that doesn't happen with a normal setup (i.e. without CAP_NET_RAW and CAP_NET_ADMIN)
1132009670 M * Bertl arok: you might want to clean up the runlevel scripts and remove network/hardware related stuff
1132009670 M * ^Cist arok, you have to adjust your init.d scripts within the vserver
1132009689 M * arok okay, thanks
1132009725 M * Pazzo ok, up and running
1132009736 M * Pazzo uname: 2.6.14.2-vs2.1.0-rc7-vs2.1-rc7-ngn0.03-g11-i686-smp
1132009751 M * daniel_hozac ouch!
1132009762 M * Pazzo hmmm...
1132009762 M * daniel_hozac i thought my unames were absurdly long.
1132009774 M * Bertl daniel_hozac: see, they aren't :)
1132009782 M * Pazzo I forgot to remove the EXTRA_NAME :(
1132009786 J * goodwill ~goodwill@netblock-66-245-192-51.dslextreme.com
1132009792 M * Bertl welcome goodwill!
1132009798 M * goodwill *blinks*
1132009799 M * Pazzo ehm - EXTRAVERSION
1132009809 M * goodwill hello
1132009812 M * Pazzo hi goodwill
1132009821 M * goodwill my you guys are friendly
1132009842 M * Bertl that's just goodwill *G*
1132009847 M * goodwill :-P
1132009853 M * Pazzo Bertl: what to do now? launch a vserver?
1132009917 M * Bertl nope, I ahve a few commands for you
1132009931 M * Pazzo ./vdevtag says "vc_ngnet_tagdev: No such process" - so the required functions should be there (as the error msg has been "function missing" or something like that before)
1132009942 M * Bertl first, use 'ip link ls' to figure the id of your lo device
1132009952 M * Pazzo eth0 -> id 2
1132009993 M * Bertl yeah, but we are looking for lo :)
1132010018 M * Pazzo ok, that's id 1
1132010026 Q * schak Ping timeout: 480 seconds
1132010035 M * Bertl so, then let's try: vdevtag -i 1 -n 1
1132010049 M * Pazzo ok
1132010058 M * Bertl and check with ifconfig or ip that lo is still there :)
1132010065 M * Pazzo still there
1132010072 M * Bertl good, now let's do
1132010089 M * Bertl vnet -C -n 666 -- tcpdump -vvnei lo &
1132010099 M * Bertl (or start it in a separate console)
1132010115 M * Bertl btw, that's the evil sniffing context ...
1132010122 M * Pazzo running in background
1132010128 M * Bertl and now let's do:
1132010134 M * Bertl vnet -C -n 42 -- ping -c 1 127.0.0.1
1132010143 M * Bertl (the friendly ping localhost)
1132010159 M * Pazzo no response, no tcpdump output
1132010173 M * Bertl ping doesn't work?
1132010196 M * Pazzo ping 127.0.0.1 works - but
1132010202 M * Pazzo vnet -C -n 42 -- ping -c 1 127.0.0.1
1132010205 M * Pazzo doesn't
1132010223 J * tonhito ~JR@200.121.194.45
1132010229 M * Bertl Pazzo: hmm, okay, unusual, but hey, the ping on the host works?
1132010236 M * Bertl welcome tonhito!
1132010256 M * Pazzo ping on the host? I can ping 127.0.0.1 from the host, yes
1132010266 M * Bertl okay, then try with:
1132010275 M * Bertl vnet -C -n 666 -- ping -c 1 127.0.0.1
1132010296 M * Pazzo Failed to create network context: File exists
1132010311 M * Bertl hmm, okay, sorry, my fault
1132010317 M * Pazzo without -C?
1132010326 M * Pazzo works
1132010329 P * tonhito 
1132010348 M * Bertl did the tcpdump record the ping now?
1132010356 M * arok okay guys, i go to bed. thanks for help. if it doesn't work i
1132010361 M * arok 'll bug you tomorrow again ;)
1132010367 M * Bertl okay, make that!
1132010378 M * arok if it works i'll tell you if you want me to
1132010383 M * Pazzo Bertl: doesn't seem so - I'll kill the process and run it on another console in foreground
1132010390 M * Bertl arok: yes sure, please do so ...
1132010399 M * arok :) good bye
1132010402 M * Pazzo tcpdump killed: 0 packets captured
1132010431 M * Bertl hmm, let me check, amybe I forgot to uplaod the latest version ... you have the build tree at hand?
1132010449 Q * arok Quit: leaving
1132010459 M * Pazzo still the same - tcpdump shows nothing
1132010464 M * Pazzo build tree is here, yes
1132010473 M * Bertl give me a second to verify ...
1132010503 M * Bertl so, in ip link ls, there is a 1: before 'lo' right?
1132010530 M * Pazzo right
1132010546 M * Pazzo btw: there is a lot of debugging output in syslog
1132010561 M * Bertl yes, we need that to narrow down the issue
1132010599 M * Bertl hmm, works flawlessly here ... checking patch now
1132010626 M * Pazzo http://pastebin.com/429709
1132010670 M * Bertl hmm, hmm ... definitely something is missing here ...
1132010883 M * Pazzo (using 2.6.14.2, patch-2.6.14.2-vs2.1.0-rc7.diff and delta-ngn0.03.diff - nothing more)
1132010894 M * Bertl yep, yep, checking ...
1132010958 M * Bertl hmm, strange ... must be a config issue, could you upload your .config file please?
1132010998 P * goodwill 
1132011004 M * Bertl and what is the output of 'vnet -C -n 99 -- ip link ls'
1132011076 M * Pazzo http://devel.gelf.net/config-2.6.14.2-vs2.1.0-rc7-vs2.1-rc7-ngn0.03-g11-i686-smp
1132011078 M * Bertl ah, I know, it's the evil_ping :)
1132011105 M * Pazzo 1: lo: <LOOPBACK,UP> mtu 16436 qdisc noqueue
1132011106 M * Pazzo  \n    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
1132011128 M * Pazzo evil_ping? heard that before from you...
1132011136 M * Bertl yeah, just a second :)
1132011145 M * Pazzo (system is debian sarge - more or less :)
1132011186 M * Bertl http://vserver.13thfloor.at/Experimental/NGNET/ping
1132011197 M * Bertl should work with the compat glibc package
1132011225 M * Bertl (if so, please try again with this one, I'll try to fix the issue in the emantime :)
1132011297 M * Pazzo ping works - tcpdump sees nothing
1132011319 M * Bertl k, now try inside context 666
1132011323 M * Pazzo (doing: ./vnet -n 666 -- ./ping -c 1 127.0.0.1)
1132011343 M * Pazzo there is 
1132011345 M * Pazzo ./vnet -C -n 666 -- tcpdump -vvnei lo
1132011346 M * Pazzo running
1132011352 M * Pazzo I did
1132011357 M * Pazzo ./vnet -n 666 -- ./ping -c 1 127.0.0.1
1132011375 M * Pazzo (./ping is your Experimental/NGNET/ping)
1132011380 M * Bertl yep
1132011384 M * Pazzo 1 packets transmitted, 1 packets received, 0% packet loss
1132011393 M * Pazzo tcpdump: 0 packets captured
1132011411 M * Bertl heh funny, so they also use different sockets for tcpdump ...
1132011422 M * Pazzo ?
1132011457 M * Bertl I also uploaded a tcpdump, try with that one
1132011485 M * Bertl daniel_hozac: do you remember what the actual difference between evil_ping and ping was?
1132011501 M * Bertl different socket types, but which ones?
1132011522 M * daniel_hozac hmm, sorry, no.
1132011533 M * Bertl np, strace will unveil that secret :)
1132011579 M * Pazzo ./tcpdump: error while loading shared libraries: libcrypto.so.0: cannot open shared object file: No such file or directory
1132011605 M * Bertl hmm, okay .. np
1132011619 M * Pazzo libgcrypt11 ?
1132011628 M * Pazzo libcrypto++5.2 ?
1132011632 M * Bertl sec
1132011661 M * Bertl rpm -qf /usr/lib/libcrypto.so.0
1132011661 M * Bertl libopenssl0-0.9.6i-1.5.82mdk
1132011744 M * Pazzo ii  openssl              0.9.7e-3
1132011787 M * Pazzo should I link /usr/lib/libcrypto.so.0.9.7 to /usr/lib/libcrypto.so.0 ?
1132011797 M * Bertl no, probably not ...
1132011810 M * Pazzo I mean: symlink
1132011813 M * Bertl let me check for the real issue and fix that, will take a few minutes
1132011823 M * Pazzo ok
1132011846 M * Bertl for the record: evil_ping does:
1132011850 M * Bertl socket(PF_INET, SOCK_RAW, IPPROTO_ICMP) = 3
1132011860 M * Bertl while 'just' ping does:
1132011865 M * Bertl socket(PF_INET, SOCK_RAW, IPPROTO_ICMP) = 3
1132011865 M * Bertl socket(PF_INET, SOCK_DGRAM, IPPROTO_IP) = 4
1132011892 M * Bertl I assume former also uses the raw socket for writing