1124410259 Q * keyser_soze Quit: Abandonando
1124413083 J * janra_ janra@paradox.homeip.net
1124413262 Q * janra Ping timeout: 480 seconds
1124414459 M * Bertl wb janra_!
1124416879 P * stefani parting (is such sweet sorrow)
1124418844 J * tchan_ ~tchan@c-24-13-81-164.hsd1.il.comcast.net
1124418943 Q * tchan Killed (NickServ command used by tchan_)
1124418956 N * tchan_ tchan
1124420653 Q * Hunger arion.oftc.net quasar.oftc.net
1124420653 Q * gregster arion.oftc.net quasar.oftc.net
1124420653 Q * Medivh arion.oftc.net quasar.oftc.net
1124420653 Q * mugwump arion.oftc.net quasar.oftc.net
1124420653 Q * dlp arion.oftc.net quasar.oftc.net
1124420653 Q * daniel_hozac arion.oftc.net quasar.oftc.net
1124420653 Q * alexx arion.oftc.net quasar.oftc.net
1124420653 Q * atsab arion.oftc.net quasar.oftc.net
1124420653 Q * BWare arion.oftc.net quasar.oftc.net
1124420653 Q * neofutur arion.oftc.net quasar.oftc.net
1124420653 Q * wibble arion.oftc.net quasar.oftc.net
1124420653 Q * pusling arion.oftc.net quasar.oftc.net
1124420653 Q * duckx arion.oftc.net quasar.oftc.net
1124420653 Q * ntrs arion.oftc.net quasar.oftc.net
1124420914 J * BWare ~bware@office.intouch.net
1124421259 J * Hunger Hunger.hu@Hunger.hu
1124421259 J * duckx ~Duck@mna75-1-81-57-39-234.fbx.proxad.net
1124421259 J * pusling pusling@195.215.29.124
1124421259 J * wibble ~tim@sophie.wobb1e.co.uk
1124421259 J * neofutur ~neofutur@neofutur.net
1124421259 J * atsab ~as@lotes.vtu.lt
1124421259 J * alexx ~alexx@proxy.ikse.net
1124421259 J * daniel_hozac ~daniel@c-6f1472d5.010-230-73746f22.cust.bredbandsbolaget.se
1124421259 J * dlp ~ddlp@tangerine.chiccp.net
1124421259 J * mugwump ~samv@watts.utsl.gen.nz
1124421259 J * Medivh ck@paradise.by.the.dashboardlight.de
1124421259 J * gregster ~gregor@greart.de
1124421300 J * ntrs ~ntrs@Dardeene-68.188.50.87.charter-stl.com
1124422059 M * Greek0 mhm. hab jetzt busybox für openwrt mit .deb support durchkompiliert, jetzt sollt' vielleicht cdebootsrap auch ohne den hack funktionieren
1124422067 M * Greek0 uff. sorry, wrong channel
1124422084 M * Bertl np
1124422108 A * Greek0 adds 'chatting' to the list of things not to do at 5.30 in the morning
1124422877 M * micah looks like util-vserver .208 with the Bertl special patches has been uploaded to debian unstable
1124422903 M * Bertl yep? great!
1124423118 M * micah yup, and our webhosting setup using vservers and syscp control panel is moving along
1124423137 M * micah I just finished a scary diagram: http://deb.riseup.net/web-server/syscp/
1124423303 M * Bertl looks interesting ..
1124424283 M * Bertl micah: do you know, does the 208+fix version include fix01 or fix02 (i.e. shiny6 or shiny7) ?
1124424359 M * micah Bertl: it appears as if it is just fix01: http://packages.qa.debian.org/u/util-vserver/news/1.html
1124424404 M * Bertl ah, well, there will be an update ... :)
1124424425 M * micah should I file a bug asking him to upload with fix02 instead?
1124424473 M * Bertl depends, with what gcc/binutils will they be typically built? if gcc 4.x, then probably yes ...
1124424523 M * micah yes gcc 4.x
1124424543 M * micah everything in debian is being transitioned to build against gcc 4.x
1124424561 M * Bertl thanks to daniel_hozac, we fixed a gcc 4 issue/bug (in the alternate syscall code) yesterday ...
1124424713 M * micah thats the fix02?
1124424718 M * Bertl sec
1124424748 M * Bertl http://vserver.13thfloor.at/Experimental/UTIL-VSERVER/
1124424811 M * Bertl http://vserver.13thfloor.at/Experimental/UTIL-VSERVER/shiny6-shiny7.diff
1124424823 M * Bertl that's the relevant part ...
1124424857 M * micah ok... I'm confused, should I ask him to upload .208 with fix02 and shiny6-shiny7.diff?
1124424898 M * Bertl no, either fix02, or, what I personally would prefer over any of the fix patches is a proper selection of the broken down changes ...
1124424903 M * Bertl in this case:
1124424909 M * Bertl http://vserver.13thfloor.at/Experimental/UTIL-VSERVER/delta-0.30.208-kheaders.diff
1124424915 M * Bertl http://vserver.13thfloor.at/Experimental/UTIL-VSERVER/delta-0.30.208-shiny7.diff
1124424920 M * Bertl and if he likes to:
1124424925 M * Bertl http://vserver.13thfloor.at/Experimental/UTIL-VSERVER/delta-0.30.208-tsfix.diff
1124424940 M * Bertl (all three together form fix02)
1124424952 M * micah ok, I agree, broken down patches are better
1124424997 M * Bertl the kheaders patch updates the kernel (vserver) headers to vs2.0
1124425045 M * Bertl the shiny patch replaces the old, broken, alternative syscall implementation by the new, extensively tested one 
1124425070 M * micah and the tsfix?
1124425106 M * Bertl the last one, the tsfix, is just a minor cosmetic change so that the test suite doesn't test uninmplemented personality stuff (at least for mandrake systems)
1124425138 M * Bertl (if that isn't an issue on debian, this patch should not be included)
1124425142 M * micah one last question -- if he applies those three diffs, he does *not* need to also apply fix01, correct?
1124425159 M * Bertl correct, fix01 is kheaders+shiny6
1124425285 M * micah I am not sure if the personality stuff is broken on debian or not
1124425293 M * micah how can I determine that?
1124425341 M * Bertl basically by compiling the package (with the tests) on various archs ppc/64, x86_64 ...
1124425386 M * Bertl if the debian build does not execute the test suite, it's probably fine without ...
1124425417 M * Bertl (but I have no idea _what_ the debian build actually does :)
1124425436 M * micah I can tell you, one second
1124425473 M * micah btw, wth presentation was very good
1124425488 M * micah but you got very easy questions from the audience :)
1124425496 M * Bertl you liked it? great!
1124425517 M * micah yes it was well organized and very clear
1124425539 M * Bertl yeah, and a walked more than a mile .. on stage :)
1124425545 M * micah hehe
1124425578 Q * Beave Ping timeout: 480 seconds
1124425580 M * micah hmm, I can't get .208 just yet
1124425587 M * micah so I'll look at .207
1124425594 M * Bertl k
1124425622 M * micah ok, it looks like it skips building on ARM because there are no reserved syscalls for that arch
1124425654 M * micah also on hppa
1124425654 M * Bertl well, that's wrong ... 
1124425657 M * micah good
1124425661 M * micah then we can fix that :)
1124425670 M * Bertl we have syscall numbers on all archs now ...
1124425675 M * micah including hppa?
1124425677 M * Bertl yep
1124425683 M * micah ok, great, I'll file a bug on those too
1124425715 M * Bertl http://linux-vserver.org/Syscall+Switch+Info
1124425722 M * micah ok, it does a configure like so:
1124425777 M * micah  ./configure --prefix=/usr --enable-release --mandir=\$${prefix}/share/man --infodir=\$${prefix}/share/info --sysconfdir=/etc --enable-dietlibc --localstatedir=/var --with-vrootdir=/var/lib/vservers
1124425819 M * micah it also passes some conffflags depending on the arch
1124425839 M * Bertl don't know about the --enable-release ...
1124425849 M * micah then does a make install install-distribution
1124425859 M * Bertl okay, so no make check then?
1124425871 M * micah so doesn't look like a make check
1124425887 M * Bertl okay, then the tsfix isn't really needed ...
1124425906 M * micah does a make all do a make check?
1124425938 M * Bertl I'd assume no ... but I'm not sure ...
1124426063 M * Bertl anyway it's probably better without the check (in the debian case) we do not want to make it break on N archs :)
1124426162 M * Bertl (the make check is basically a developer test to see if everything is working as expected, which works fine on tested systems and on the target hardware of course, but will fail for crossbuilds and such)
1124426430 M * Bertl okay folks ... I'm off to bed now ... cya later ...
1124426469 N * Bertl Bertl_zZ
1124427193 J * Beave ~beave@vistech.org
1124431479 Q * saddi Quit: using sirc version 2.211+KSIRC/1.3.12
1124432618 J * maharaja maharaja@80.64.143.52
1124433292 Q * Hunger Ping timeout: 480 seconds
1124433789 J * Hunger Hunger.hu@Hunger.hu
1124433895 J * sven2 ~sven@GKCDVI.dsl.saunalahti.fi
1124443852 Q * Hunger helium.oftc.net iridium.oftc.net
1124443852 Q * ntrs helium.oftc.net iridium.oftc.net
1124443852 Q * gregster helium.oftc.net iridium.oftc.net
1124443852 Q * Medivh helium.oftc.net iridium.oftc.net
1124443852 Q * mugwump helium.oftc.net iridium.oftc.net
1124443852 Q * dlp helium.oftc.net iridium.oftc.net
1124443852 Q * daniel_hozac helium.oftc.net iridium.oftc.net
1124443852 Q * alexx helium.oftc.net iridium.oftc.net
1124443852 Q * atsab helium.oftc.net iridium.oftc.net
1124443852 Q * wibble helium.oftc.net iridium.oftc.net
1124443852 Q * duckx helium.oftc.net iridium.oftc.net
1124443852 Q * pusling helium.oftc.net iridium.oftc.net
1124443852 Q * neofutur helium.oftc.net iridium.oftc.net
1124443852 Q * Beave helium.oftc.net iridium.oftc.net
1124443852 Q * BWare helium.oftc.net iridium.oftc.net
1124443852 Q * janra_ helium.oftc.net iridium.oftc.net
1124443852 Q * renihs helium.oftc.net iridium.oftc.net
1124443852 Q * lonewolff helium.oftc.net iridium.oftc.net
1124443853 Q * Loki|muh helium.oftc.net iridium.oftc.net
1124443853 Q * lilo helium.oftc.net iridium.oftc.net
1124443853 Q * Vudumen helium.oftc.net iridium.oftc.net
1124443853 Q * revenger helium.oftc.net iridium.oftc.net
1124443853 Q * sladen helium.oftc.net iridium.oftc.net
1124443853 Q * case helium.oftc.net iridium.oftc.net
1124443853 Q * dsoul helium.oftc.net iridium.oftc.net
1124443853 Q * tchan helium.oftc.net iridium.oftc.net
1124443853 Q * jkl_ helium.oftc.net iridium.oftc.net
1124443853 Q * logger helium.oftc.net iridium.oftc.net
1124443853 Q * DuckMaster helium.oftc.net iridium.oftc.net
1124443853 Q * skceb helium.oftc.net iridium.oftc.net
1124443853 Q * BobR_oO helium.oftc.net iridium.oftc.net
1124443853 Q * virtuoso helium.oftc.net iridium.oftc.net
1124443853 Q * chand helium.oftc.net iridium.oftc.net
1124443853 Q * locksy helium.oftc.net iridium.oftc.net
1124443853 Q * derbien helium.oftc.net iridium.oftc.net
1124443853 Q * sid3windr helium.oftc.net iridium.oftc.net
1124443853 Q * ray6 helium.oftc.net iridium.oftc.net
1124443854 Q * Bertl_zZ helium.oftc.net iridium.oftc.net
1124443854 Q * zobel helium.oftc.net iridium.oftc.net
1124443854 Q * obi helium.oftc.net iridium.oftc.net
1124443944 J * Hunger Hunger.hu@Hunger.hu
1124443944 J * Beave ~beave@vistech.org
1124443944 J * ntrs ~ntrs@Dardeene-68.188.50.87.charter-stl.com
1124443944 J * gregster ~gregor@greart.de
1124443944 J * Medivh ck@paradise.by.the.dashboardlight.de
1124443944 J * mugwump ~samv@watts.utsl.gen.nz
1124443944 J * dlp ~ddlp@tangerine.chiccp.net
1124443944 J * daniel_hozac ~daniel@c-6f1472d5.010-230-73746f22.cust.bredbandsbolaget.se
1124443944 J * alexx ~alexx@proxy.ikse.net
1124443944 J * atsab ~as@lotes.vtu.lt
1124443944 J * neofutur ~neofutur@neofutur.net
1124443944 J * wibble ~tim@sophie.wobb1e.co.uk
1124443944 J * pusling pusling@195.215.29.124
1124443944 J * duckx ~Duck@mna75-1-81-57-39-234.fbx.proxad.net
1124443944 J * BWare ~bware@office.intouch.net
1124443944 J * tchan ~tchan@c-24-13-81-164.hsd1.il.comcast.net
1124443944 J * janra_ janra@paradox.homeip.net
1124443944 J * renihs ~renihs___@193.170.52.70
1124443944 J * zobel zobel@zobel.irc.ftbfs.de
1124443944 J * obi ~obi@asus.saftware.de
1124443944 J * lonewolff ~lonewolff@host86-128-17-74.range86-128.btcentralplus.com
1124443944 J * Loki|muh loki@satanix.de
1124443944 J * lilo ~lilo@lilo.usercloak.oftc.net
1124443944 J * case ~case@donpanic.faveve.uni-stuttgart.de
1124443944 J * sladen paul@starsky.19inch.net
1124443944 J * dsoul darksoul@vice.ii.uj.edu.pl
1124443944 J * revenger ~joe@bulldog.infosys.de
1124443944 J * Vudumen vudumen@perverz.hu
1124443944 J * DuckMaster ~duckx@195.75.27.158
1124443944 J * logger ~rs@vds.pas-mal.com
1124443944 J * jkl_ eric@c-67-190-142-149.hsd1.co.comcast.net
1124443944 J * ray6 ~ray@klon3.gcsc2.ray.net
1124443944 J * sid3windr luser@bastard-operator.from-hell.be
1124443944 J * derbien ~derbien@whiterabbit.nbmc.de
1124443944 J * locksy ~locksy@mrtg.sisgroup.com.au
1124443944 J * chand ~chand@staff.lycos.fr
1124443944 J * virtuoso ~s0t0na@80.253.205.251
1124443944 J * BobR_oO ~georg@212.16.62.52
1124443944 J * Bertl_zZ ~herbert@212.16.62.52
1124443944 J * skceb skceb@bang-bang.feuer-frei.com
1124446762 Q * sven2 Ping timeout: 480 seconds
1124447574 J * Pazzo ~Pazzo@host130-250.pool8172.interbusiness.it
1124447948 N * pg|cereal pg`aw|cereal
1124448299 M * wibble hmm
1124448317 M * wibble can I specify vserver build with the --nodev option, but still bind it to an existing ip?
1124448328 M * wibble like --nodev --ip 10.0.0.254 ?
1124448417 M * Pazzo should work, yes - the ip will not be assigned/removed at vServer's start and stop actions, you will have to add it manually
1124448431 M * wibble cool, will give that a go
1124448482 M * Pazzo I'm not sure about the create syntax, I normally created it using also the device as a parameter and then modified /etc/vservers/*
1124448514 M * wibble yeah I can touch /etc/vservers/<host>/interface/0/nodev 
1124448515 M * wibble which works
1124448527 M * wibble as there is a problem with vlan interfaces in the current tools
1124448799 M * wibble hmm
1124448823 M * wibble d0h
1124448853 M * Pazzo (back later)
1124448855 M * wibble cya
1124448861 N * Pazzo PazZzzooo
1124450392 M * BWare wibble: what version of the tools ?
1124450401 M * wibble 0.30.208
1124450415 M * BWare I am running .204 and have no problems with vlans
1124450425 M * BWare I haven't tested .208 yet
1124450448 M * wibble it basically brings up the vlan interface with a 127.0.0.1 address and doesn't destroy the vlan interface correctly
1124450466 M * wibble i believe the destroy is fairly trivial to fix, but cant see what the problem is with the loopback addy
1124450466 M * BWare You are not using named aliases ?
1124450476 M * wibble ?
1124450495 M * wibble http://savannah.nongnu.org/bugs/?func=detailitem&item_id=14148
1124450499 M * BWare . /etc/vservers/NAME/interfaces/ID/name
1124450602 M * BWare That creates a named alias of ethX.VLANID and assigns the correct ip to the alias
1124450645 M * BWare Gives an additional interface in the list, but it should work ;)
1124450690 M * wibble yeah, my workaround is to touch /etc/vservers/NAME/interfaces/ID/nodev
1124450698 M * wibble and configure the ip interface manually during boot
1124450703 M * wibble works just fine for me..
1124450723 M * wibble but the vserver tools should bring up the interface ok, which it doesn't
1124450739 M * wibble the prob is also in 207
1124451053 M * BWare oke... I'll stay away from those ;)
1124451108 M * wibble heh, the only thing I have found is the vlan issue, everything else seams just fine"
1124451471 M * BWare Well I am using lots of vlan based vservers, so if that's the only issue, it's still a no go for me
1124451546 Q * lonewolff Quit: leaving
1124452118 J * lonewolff ~lonewolff@host86-128-17-74.range86-128.btcentralplus.com
1124452118 Q * Doener Read error: Connection reset by peer
1124452241 J * Doener ~doener@p54877B68.dip.t-dialin.net
1124452403 M * wibble BWare: do you use iproute2 with multiple gateways for your vlan'ed vservers?
1124452434 J * erwan_taf ~erwan@81.80.43.77
1124452437 P * erwan_taf 
1124452572 M * BWare wibble: yep
1124456533 Q * cryo Remote host closed the connection
1124457003 J * sven2 ~sven@GMMDCCCVI.dsl.saunalahti.fi
1124457114 J * cryo ~say@212.86.243.154
1124457449 Q * sven2 Remote host closed the connection
1124457940 M * wibble Bertl_zZ: I think we have probs with that reiser patch, or something funky is going on!
1124461076 N * Bertl_zZ Bertl
1124461095 M * Bertl morning folks!
1124461105 M * Bertl wibble: what kind of probs?
1124464113 N * Bertl Bertl_oO
1124465173 Q * lownoize Ping timeout: 480 seconds
1124465316 J * stefani ~stefani@superquan.apl.washington.edu
1124465781 J * lownoize ~lownoize@p54AC9AE9.dip0.t-ipconnect.de
1124466775 M * Doener Bertl_oO: hm, telling from your wget "screenshot" the compressed files don't have a Content-size(length?) header. Some programs seem to have problems with that (for example I've seen IE fail miserably on such downloads in some cases)  
1124466828 M * Doener but as you're probably dealing with on-the-fly compression there, i have no idea whether the size can be sent before-hand anyways... 
1124468598 Q * obi arion.oftc.net plasma.oftc.net
1124468598 Q * zobel arion.oftc.net plasma.oftc.net
1124468837 Q * sid3windr Ping timeout: 480 seconds
1124468898 J * zobel zobel@zobel.irc.ftbfs.de
1124468898 J * obi ~obi@asus.saftware.de
1124469168 J * sid3windr luser@bastard-operator.from-hell.be
1124469286 Q * BWare Ping timeout: 480 seconds
1124469690 Q * DuckMaster Quit: Leaving
1124469730 N * Bertl_oO Bertl
1124469773 M * Bertl Doener: correct, but that is quite fine in http, and I usually store the compressed (bz2) version ...
1124470208 J * BWare ~bware@office.intouch.net
1124470576 M * Doener Bertl: yeah, just talking 'bout broken clients ;)
1124470684 M * Bertl I know, was just for the record :)
1124470724 M * Doener k :)
1124471906 M * Hollow hey folks
1124472410 M * Bertl hey Hollow!
1124472473 M * Hollow hey Bertl, saw the loopback fix.. what is it about?
1124472485 M * Matthew-1 hello everyone
1124472523 M * Matthew-1 i got an issue with trying to use vserver in a LVS cluster
1124472544 M * Matthew-1 i'm trying to use direct route
1124472577 M * Matthew-1 i don't if vserver supports this kind of thing
1124472910 M * Bertl Hollow: we are not sure yet, but it sounds (and looks) reasonable :)
1124472943 J * sven111 ~sven@GMMDCCCVI.dsl.saunalahti.fi
1124472947 M * Hollow lol
1124472950 M * Bertl Matthew-1: linux-vserver should support nearly all that an unpatched kernel supports ...
1124472955 M * Bertl welcome sven111!
1124472967 M * sven111 thanks :)
1124473074 M * sven111 I got a vserver running on Arch linux, but I am not sure how to get each vserver connected to the net
1124473103 M * sven111 I'd like to make apache run on each vserver
1124473111 M * Bertl ah, did you update the wiki pages?
1124473117 M * sven111 yes
1124473126 M * Bertl I was just reading about Arch linux ...
1124473138 M * sven111 I got advice here before
1124473165 M * Bertl and I'm pretty sure you will get today too ...
1124473191 M * Bertl so what is your network setup?
1124473211 M * sven111 so is it possible to somehow set up a kind of a virtual nat between vservers and host?
1124473256 M * sven111 I've been just testing things for now so I have just one ip from dhcp
1124473266 M * sven111 on host
1124473268 M * Bertl virtual nat ... hmm, you can give each guest a private IP and use S/DNAT to use/share a public one ... if that is what you mean ...
1124473332 M * sven111 I've been trying to give ip like 192.168.0.x but then checked in a vhost and it showed no ip at all
1124473353 M * Bertl showed as in 'ifconfig not showing anything'?
1124473386 M * Matthew-1 but does it support direct route LVS,  because for that to work i have to have a loopback interface on the host with the IP of the vitrual LVS ip
1124473415 M * Matthew-1 so how would i add another loopback interface in the vserver
1124473423 M * sven111 just inet addr part is missing
1124473423 M * Bertl Matthew-1: you have all kinds of interfaces (the ones you also have on vanilla kernels) on the host ...
1124473434 M * Matthew-1 oh
1124473494 M * Bertl sven111: I just read about 'Modern Tools' on the arch linux page ... iproute2 is at least around since 1999 (that's how far their archives go back), give the 'ip' tool a try on your guest ...
1124473498 M * Matthew-1 ok i see now
1124473509 M * Matthew-1 i'll try it
1124473510 M * Matthew-1 thanks
1124473515 M * Bertl yw
1124473550 A * Hollow wonders why ifconfig is still shipped with the base system (at least in gentoo land)
1124473578 M * Hollow i don't need it to run my system, ergo make it optional
1124473590 M * Hollow (that would be the gentoo way ;)
1124473638 M * daniel_hozac people are too used to it ;)
1124473644 M * Hollow damn ;)
1124473692 M * renihs ya ifconfig is still included, and i am not sure if iproute2 is too per default
1124473704 M * Hollow in gentoo it isn't
1124473709 M * renihs pff
1124473715 M * renihs it should be default since 1995
1124473727 M * sven111 thanks,Bertl,I'll see the page too then
1124473732 M * Doener at least you can use the iproute stuff with the gentoo network setup stuff now
1124473750 M * renihs i never used the gentoo network stuff, cant figure it out :)
1124473777 M * Bertl sven111: yw, keep in mind, S/DNAT not MASQUERADE ...
1124473779 M * Hollow well, the net-tools ebuild (which includes ifconfig) also includes other apps like hostname and netstat
1124473793 M * renihs yah, MASQUERADE , me is still slamming his head....
1124473802 M * renihs how should it work without input device :)
1124473812 M * Bertl sven111: and if you feel more comfortable with ifconfig showing something, use an alias or separate interface (e.g. dummy0)
1124473826 M * Doener Hollow: iproute's "ss" replaces netstat ;)
1124473839 M * Hollow ah
1124473863 M * Doener needs some kernel support in the kernel network support, don't remember which option it was
1124473897 M * sven111 Bertl: ok,I'll try to read more about these things
1124473908 M * Hollow reminds me... do i need to have CONFIG_IP_TCPDIAG for ss?
1124473923 M * Hollow ok
1124473925 M * Hollow heh
1124474487 Q * virtuoso Ping timeout: 480 seconds
1124474536 M * renihs gruml, i doubt this is a vserver related issue, but has anyone ever experienced problems detaching loop devices? (cant free /dev/loop/<num>) after using them
1124474571 M * Bertl maybe they are still mounted?
1124474674 M * renihs /proc/mounts shows nothing
1124474682 M * Bertl maybe inside a guest which was started after the loop mount?
1124474694 M * renihs nope i did the loopmount from the host system
1124474699 J * virtuoso ~s0t0na@80.253.205.251
1124474699 M * renihs losetup /dev/loop7 shows
1124474704 M * renihs /dev/loop7: [0801]:699 (/vservers/backup/2005-08-19/_mta)
1124474717 M * renihs but the file is unmounted and doesnt exist anymore
1124474733 M * Bertl in the host namespace, yes, but in the guests?
1124474744 M * renihs unmounted and losetup -d correctly through a script
1124474748 M * renihs hmm 
1124474757 M * renihs it was never mounted in the guest
1124474772 M * Bertl but maybe a guest was started _after_ it was mounted on the host?
1124474776 M * renihs only from host -> copied all files (from a guest) to a loopback image 
1124474783 M * renihs hmm yes
1124474790 M * renihs the script stops and starts the vserver
1124474796 M * Bertl then that one will have inherited the loop mount
1124474826 M * renihs u mean, if i have a loopmount on the host, if i restart a guest it will inherit the mount?
1124474849 M * Bertl yes, because the guest makes a copy of your hosts namespace ...
1124474861 M * renihs k, i will reorder the script
1124474884 M * Bertl you can use vnamespace to enter the guest's namespace ...
1124474901 M * Bertl there you should see the mount in /proc/mounts and probably can unmount it ...
1124474921 M * Bertl of course, you can also stop and start the guest
1124474970 M * renihs 1) loopmount empty image -> rsync files from /vserver/bla to empty image -> stop vserver bla -> rsync files again -> start vserver bla again 
1124474975 M * renihs that is my order at the moment
1124475007 M * renihs cant see the mount in the guest however i will detach/retach the loop before and after restarting the vserver i guess
1124475116 M * Bertl the mount is not in the filesystem of the guest, but in the guest's namespace
1124475142 M * renihs k
1124475982 M * renihs hmm one thing i learned in this channel is for sure :)
1124475988 M * renihs Bertl is *always* right :)
1124475998 M * renihs thx again :)
1124476041 M * Bertl you're welcome! and no, I'm not _always_ right ...
1124476051 M * renihs until now, yes you are :)
1124476082 M * renihs and my total vserver downtime for backup is 7 sec
1124476088 M * renihs thats ok i guess
1124476184 M * Bertl interesting, you backup via loopback devices?
1124476228 M * renihs yap, i create an image with ext3/reiser with 30% more space than actual use from vserver -> rsync everything to image and compress it
1124476244 M * renihs that way i only need to decompress the image and loopmount it in case of recovery
1124476260 M * renihs maybe not elegant but seems to work
1124476281 M * Bertl interesting approach ...
1124476306 M * renihs "interesting" is a good word
1124476309 M * renihs can mean alot :)
1124476311 M * Bertl you might want to look into LVM snapshots in the future ...
1124476315 M * renihs yah i know
1124476323 M * renihs lvm2 can make snapshots from running system i heard
1124476336 M * renihs didnt touch lvm for a long time :(
1124476352 M * eyck lvm used to be able to do that
1124476366 M * eyck xfs does that on filesystem level,
1124476370 M * renihs ah, good to know, another thing i never tried out
1124476398 M * renihs hmm so with xfs i could also make a snapshot of a running database server for example?
1124476435 M * Bertl I doubt it ...
1124476441 M * renihs k :)
1124476457 M * Bertl eyck: do you know the xfs folks?
1124476457 M * eyck running databases are notoriously hard to snapshot,
1124476472 M * eyck neither xfs nor lvm will help you with that,
1124476481 M * eyck nope, I'm not very sociable person you know,
1124476503 M * Bertl well, you could have bugged them some time ago, no? :)
1124476505 M * renihs k, but my approach should work i guess? rsync (dirty) from a running server all files, shut down server, rsync (clean) all files which changed again), start vserver?
1124476527 M * Bertl renihs: yeah, probably a good approach ...
1124476540 M * eyck yupp, that's quite possible, I think I bugged them over something few times..
1124476544 M * renihs k, because the 2nd rsync is really fast
1124476552 M * eyck oh yeah, software RAID issues,
1124476592 M * Bertl because I'd like to talk to some xfs developer, because I'd like to add some xfs stuff (vserver specific)
1124476594 M * eyck because software raid uses it's own superblock, and it's kept at the same place as xfs's superblock,
1124476622 M * eyck ext2/ext3 is smart enough about this to have a hole where software raid superblock is supposed to be
1124476647 M * eyck hmm, don't tell them you know me then...
1124476648 M * Bertl always thought the raid superblock is outside the raid volume? 
1124476658 M * renihs me too
1124476665 M * renihs at least i hope so?
1124476671 M * eyck you know, they weren't very responsive to my suggestions... ;) 
1124476681 M * eyck what do you mean 'outside' ?
1124476701 M * Bertl |----------- filesystem -------------| Raid SB |
1124476712 M * eyck  say, you've got /dev/sda5 and /dev/sda6, and build /dev/md1 out of that...
1124476734 M * eyck then on /dev/sda5 and sdb6 you get superblocks, those would be invisible from /dev/md1, 
1124476737 M * renihs sda5 and sdb5 i guess?
1124476744 M * eyck yeah, right, sorry,
1124476747 M * Bertl ithen I'd suspect /dev/md1 to be slightly smaller than whatever combination of sda5/6
1124476777 M * Bertl i.e. precisely smaller by that two superblocks :)
1124476784 M * eyck exactly, but ext2/ext3 knows about this.. and you can use ext2/3 filesystem built on /dev/md1 from /dev/sda5
1124476798 M * eyck and the other way round also, very cool
1124476825 M * Bertl so you are talking about a RAID-1, you acces via one of the raid sources, yes?
1124476847 M * eyck  xfs developers don't like such tricks though, they don't like unreliable hardware too, 
1124476869 M * Bertl because I don't think that a raid-5 or stripe set will make you happy ...
1124476891 M * eyck yeah, but then again, there are no reasons for sane person to use raid5.
1124476920 M * Bertl aha? probably I'm insane then ... ;)
1124476934 M * eyck hmm, just young probably,
1124476949 M * Bertl so why do you think raid5 is evil?
1124476954 M * eyck few more years and wisdom will dawn on you 
1124476990 M * eyck  just useless,
1124477015 M * Bertl please enlighten me, why is raid5 over 4 disks useless?
1124477067 M * eyck please enlighten me, why would you raid5 four disks, instead of raid10 or two raid1s?
1124477086 M * Bertl because of the disk space and the performance?
1124477115 M * eyck performance? IE you want your disks to be slow? why?
1124477127 M * eyck what do you need disk space for?
1124477149 M * eyck you need more space, you buy bigger disks,
1124477160 M * renihs raid5 is that much slower?
1124477170 M * eyck or more disks if performance is critical
1124477174 M * Bertl hehe, if you think raid5 is slower than raid10 or linear raid1, you should do your math ...
1124477193 M * eyck maybe I should go study math then,
1124477204 M * eyck I heard that math is hard though,
1124477205 M * renihs raid1 is faster to read i guess, slower to write, raid5 is slow on both?
1124477248 M * Bertl renihs: no, 4 disk raid5 will be able to read from 3 disks at once and/or read at 3 separate locations
1124477250 M * eyck I've never met a DBA who wouldn't shot me for setting up raid5 for him.
1124477281 M * Bertl renihs: of course, the write path is more expensive than raid1 (but only slightly)
1124477306 M * renihs sounds reasonable
1124477317 M * eyck plus, every single RAID5 I've had broke down,
1124477340 M * renihs i only used raid1?s aggregated with lvm :)
1124477342 M * eyck there was this one that defied that trend,
1124477347 M * Bertl even raid10 (you do not want raid01) isn't that performant if done in software
1124477351 M * eyck but it broke last month
1124477376 M * eyck so there, 100% failure rate for raid5.
1124477410 M * Bertl renihs: but of course raid10 is better from the fail case scenario ...
1124477422 M * Bertl (but also more expensive, as you get less disk space)
1124478093 M * renihs gruml, now my first loop device failes to unloop :)
1124478136 M * renihs ververs are running, setup loop device (losetup) from host system, rsync files from /vservers/bla to loopback image -> detach loopback failes
1124478138 M * renihs narf
1124478183 M * Bertl hmm, you ahve a mount/umount there you didn't mention, right?
1124478196 M * renihs umount and losetup -d 
1124478202 M * renihs and losetup and mount yes
1124478340 M * renihs http://www.doebling.sth.ac.at/livedvd/backup.log shows the steps, 
1124478355 M * renihs but i guess i am trying to figure this one out when my gf releases me from her claws
1124478410 M * Bertl *claw* *claw* :)
1124478454 M * renihs ga, i spend at least 30% of my non sleeping time with her, and its still not enough
1124478476 M * renihs i was feeding ducks all yesterday afternoon with her, in an romantic park picnick
1124478487 M * Bertl sounds great!
1124478497 M * renihs hmm
1124478507 M * renihs the ducks where great, i have to admit
1124478543 M * renihs beeing a duck in such a park, beeing fed by ppl isnt too bad i guess
1124478551 M * renihs but now i gotta go, cya later
1124478594 M * Bertl cya, and send greetings!
1124479361 M * eyck yeah, because nothing would make her happier than reciving greetings from your geeky friends
1124479686 M * Bertl especially from eyck, of course :)
1124479718 M * Bertl wibble: do you have any details for me?
1124479876 M * Bertl okay, np, will be off for dinner (and movie), will be back later ...
1124479909 N * Bertl Bertl_oO
1124481190 M * eyck what is wrong with you people...dinners...movies...GIRLFRIENDS ?!
1124482410 M * Hollow off to bed, cu tomorrow (or probably not ;)
1124483453 M * Matthew-1 how would i set up a loopback interface inside of vserver?
1124483466 M * Matthew-1 i'm using version 2.0
1124483563 M * Matthew-1 btw it's a virtual loopback interface for an LVS cluster
1124486031 M * renihs ha, now my stupid backup script works, thx alot bertl!
1124486040 M * renihs http://www.doebling.sth.ac.at/vserver/backup-test.log
1124486134 M * lownoize looks nice ;)
1124486150 M * renihs the times are ok for an old 1200 mhz athlon something 
1124486181 M * renihs output is ugly, need to work on that too i guess :), but who cares 
1124486200 M * lownoize you make a backup too a loopback mounted filesystem?
1124486389 M * renihs yah 
1124486390 M * renihs faster when recovering, i only need to unzip the image and loopmount it 
1124486390 M * renihs than when i have to copy 1000nds of files
1124486390 M * renihs 54869 to be precise
1124486390 M * renihs i could easily make everything much smaller, but too lazy for that
1124486390 M * renihs but its a quick & cheap solution this way i guess, didnt have time to read up on lvm or anything fancy
1124486392 J * Aiken ~james@tooax6-103.dialup.optusnet.com.au
1124486417 M * lownoize sounds interesting
1124486468 M * renihs dunno, i dont actually need those backups, only need (etc and var maybe) but i thought if i already have a complete hardware independet vserver setup, creating backups might be a good idea
1124486514 M * renihs its just a router / smtp / dns / some web and other stuff
1124486522 M * renihs no important databases etc, student home :)
1124486566 M * lownoize hehe
1124486750 M * renihs hmm i am really the closest living relative to lemming if at all...
1124486787 M * renihs reading my own writings make me remember that my great backup script backups everything except the /etc/vserver/<bla> configuration files ...
1124486793 M * lownoize i make backups with tar to a tapedrive
1124486809 M * renihs bla, i never ever had a tapedriver :((
1124486827 M * renihs i still believe that is very convienient 
1124486832 M * renihs (keep the typos)
1124486872 M * lownoize but only once a month or something
1124486956 M * renihs would be more than enough in my szenarios
1124486971 M * lownoize same here it's ok for home use
1124486974 M * renihs once a week, and delete all ones i am planning
1124486981 M * renihs old ones i mean
1124486982 M * renihs bla :)
1124487027 M * renihs if its done via cron, i only have to think about it once :)
1124489107 Q * renihs Remote host closed the connection
1124490595 Q * zobel Ping timeout: 480 seconds
1124490977 J * zobel zobel@zobel.irc.ftbfs.de
1124491300 J * mnemoc ~amery@200.75.27.30
1124491393 Q * mnemoc Quit: Changing server
1124491505 J * mnemoc ~amery@200.75.27.30
1124492200 N * Bertl_oO Bertl
1124492215 M * Bertl evening mnemoc!
1124492288 M * mnemoc hi Bertl :)
1124492312 J * renihs ~renihs___@193.170.52.70
1124492335 M * Bertl mnemoc: so you are looking for a backport, right?
1124492344 M * Bertl wb renihs!
1124492374 M * mnemoc Bertl: i could live with rc4 but a backport of 2.0 would be nice :)
1124492413 M * renihs hi Bertl!
1124492439 M * Bertl mnemoc: well, let's see what we changed since rc4 (that works for you, right?)
1124492456 M * mnemoc Bertl: yes, i have rc4 working :)
1124492483 M * Bertl mnemoc: are you going to maintain it? i.e. update when we update the kernel?
1124492521 M * mnemoc Bertl: i'm maintaining t2 stable branch... i'll only update kernel when all the kernel extensions we support get updated :)
1124492536 M * mnemoc i'm fine staying at 2.6.11 anyway :)
1124492560 M * Bertl no, what I meant is: if we release 2.0.1 (with a bunch of fixes) will you update to that?
1124492566 M * mnemoc sure
1124492651 M * mnemoc updates on rock were a big bottleneck, but not on t2.... that's one of the reasons of the fork
1124492675 M * Bertl okay, sounds good to me ... let's see ...
1124493283 M * Bertl mnemoc: want some cake or some tea?
1124493345 M * mnemoc Bertl: cake is good :) tghanks
1124493353 M * mnemoc s/good/fine/ :p
1124493452 A * Bertl hands over a piece of cake ...
1124493570 M * mnemoc looks tasteful, thanks
1124495953 Q * Doener Ping timeout: 480 seconds
1124495961 J * Doener ~doener@p548769FC.dip.t-dialin.net
1124495973 M * Bertl wb Doener!