1156291691 J * ntrs__ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com
1156291691 Q * ntrs_ Read error: Connection reset by peer
1156292809 J * cehteh ~ct@cehteh.homeunix.org
1156293827 J * ntrs_ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com
1156293827 Q * ntrs__ Read error: Connection reset by peer
1156294335 J * ntrs__ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com
1156294336 Q * ntrs_ Read error: Connection reset by peer
1156294962 J * ntrs ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com
1156294962 Q * ntrs__ Read error: Connection reset by peer
1156295241 J * ntrs_ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com
1156295241 Q * ntrs Read error: Connection reset by peer
1156295792 Q * coocoon Ping timeout: 480 seconds
1156296511 J * mire ~mire@110-167-222-85.COOL.ADSL.VLine.Verat.NET
1156296581 J * coocoon ~coocoon@p54A07AEE.dip.t-dialin.net
1156296959 J * ntrs__ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com
1156296960 Q * ntrs_ Read error: Connection reset by peer
1156298267 Q * gerrit_ Ping timeout: 480 seconds
1156299482 Q * wenchien Ping timeout: 480 seconds
1156299488 J * jesse_ ~wenchien@221-169-69-23.adsl.static.seed.net.tw
1156299512 N * jesse_ wenchien
1156302904 J * abi` ~abi@enz.schiach.de
1156302988 Q * abi Read error: Connection reset by peer
1156304360 J * undefined ~undefined@adsl-68-93-106-76.dsl.rcsntx.swbell.net
1156304437 M * undefined could somebody please remind me where in the source the maximum number of ip addresses is set for all vservers?
1156304820 M * undefined is it NB_IPV4ROOT?
1156304829 M * lylix in the kernel source it is:
1156304872 M * lylix kernel/vserver/network.h
1156304900 M * undefined include/linux/vserver/network.h: #define NB_IPV4ROOT    16
1156304936 M * undefined at least that looks most likely it from my browsing of the vserver kernel patch
1156304943 M * lylix oh hold on
1156304985 M * undefined especially when i go finding statements in hte code like this:
1156304987 M * lylix network.h
1156304997 M * undefined int vc_set_ipv4root(uint32_t nbip, void __user *data)
1156305007 M * undefined +       if (nbip < 0 || nbip > NB_IPV4ROOT)
1156305007 M * undefined +               return -EINVAL;
1156305053 M * lylix let me get the full path
1156305108 M * lylix -> /usr/src/linux/include/linux/vserver/network.h
1156305128 M * lylix assuming 'linux' is symlinked to your kernel source tree
1156305192 M * lylix so your browsing was correct
1156305794 M * undefined lylix: thanks!
1156307391 Q * coocoon Quit: KVIrc 3.2.0 'Realia'
1156308428 J * coocoon ~coocoon@p54A06219.dip.t-dialin.net
1156308534 Q * FloodServ charon.oftc.net oxygen.oftc.net
1156308534 Q * Greek0 charon.oftc.net oxygen.oftc.net
1156308534 Q * Nam charon.oftc.net oxygen.oftc.net
1156308534 Q * shedi charon.oftc.net oxygen.oftc.net
1156308534 Q * Johnsie charon.oftc.net oxygen.oftc.net
1156308534 Q * micah charon.oftc.net oxygen.oftc.net
1156308534 Q * tanjix charon.oftc.net oxygen.oftc.net
1156308534 Q * mugwump charon.oftc.net oxygen.oftc.net
1156308534 Q * harry charon.oftc.net oxygen.oftc.net
1156308534 Q * abi` charon.oftc.net oxygen.oftc.net
1156308534 Q * wenchien charon.oftc.net oxygen.oftc.net
1156308534 Q * ntrs__ charon.oftc.net oxygen.oftc.net
1156308534 Q * Aiken charon.oftc.net oxygen.oftc.net
1156308534 Q * JimmyGulp charon.oftc.net oxygen.oftc.net
1156308534 Q * glutoman charon.oftc.net oxygen.oftc.net
1156308534 Q * DreamerC charon.oftc.net oxygen.oftc.net
1156308534 Q * zkbrsnie charon.oftc.net oxygen.oftc.net
1156308534 Q * rob-84x^ charon.oftc.net oxygen.oftc.net
1156308534 Q * hap charon.oftc.net oxygen.oftc.net
1156308534 Q * [PUPPETS]Gonzo charon.oftc.net oxygen.oftc.net
1156308534 Q * tokkee charon.oftc.net oxygen.oftc.net
1156308534 Q * mcp charon.oftc.net oxygen.oftc.net
1156308534 Q * nokoya charon.oftc.net oxygen.oftc.net
1156308557 J * abi` ~abi@enz.schiach.de
1156308557 J * wenchien ~wenchien@221-169-69-23.adsl.static.seed.net.tw
1156308557 J * ntrs__ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com
1156308557 J * Greek0 ~greek0@85.255.145.201
1156308557 J * Nam ~nam@S0106001195551ff0.va.shawcable.net
1156308557 J * Aiken ~james@tooax6-237.dialup.optusnet.com.au
1156308557 J * shedi ~siggi@inferno.lhi.is
1156308557 J * JimmyGulp ~james@ns0.esagroup.co.uk
1156308557 J * glutoman glut@no.suid.pl
1156308557 J * Johnsie ~jdlewis@static-acs-24-154-32-33.zoominternet.net
1156308557 J * micah ~micah@micah.riseup.net
1156308557 J * tanjix ~tanjix@office.star-hosting.de
1156308557 J * DreamerC ~dreamerc@59-112-6-184.dynamic.hinet.net
1156308557 J * zkbrsnie ~zkbrsnie@83-64-146-226.klosterneuburg.xdsl-line.inode.at
1156308557 J * rob-84x^ rob@submarine.ath.cx
1156308557 J * mugwump ~samv@watts.utsl.gen.nz
1156308557 J * nokoya young@hi-230-82.tm.net.org.my
1156308557 J * mcp ~hightower@wolk-project.de
1156308557 J * tokkee tokkee@casella.verplant.org
1156308557 J * hap ~penso@212.27.33.226
1156308557 J * [PUPPETS]Gonzo gonzo@langweiligneutral.deswahnsinns.de
1156308557 J * harry ~harry@d54C2508C.access.telenet.be
1156308557 J * FloodServ services@services.oftc.net
1156309755 Q * FireEgl Quit: Bye...
1156310239 Q * coocoon Quit: KVIrc 3.2.0 'Realia'
1156310356 J * s0undt3ch_ yfrtafan@bl7-241-134.dsl.telepac.pt
1156310427 Q * balbir Quit: Leaving
1156310436 J * coocoon ~coocoon@p54A06219.dip.t-dialin.net
1156310779 Q * s0undt3ch Ping timeout: 480 seconds
1156310779 N * s0undt3ch_ s0undt3ch
1156311018 J * gerrit_ ~gerrit@c-67-160-146-170.hsd1.or.comcast.net
1156313885 M * undefined hmmm, it appears that increasing NB_IPV4ROOT is not necessary for my situation
1156313935 M * undefined as i have more than 16 vserver guests, but only 2 IPs per vserver guest
1156313955 M * doener yeah, the limit is per vserver
1156313981 M * undefined doener: thanks for the confirmation
1156314019 P * undefined 
1156314211 Q * gerrit_ Ping timeout: 480 seconds
1156314431 Q * Johnsie Ping timeout: 480 seconds
1156314439 Q * lylix Ping timeout: 480 seconds
1156314934 J * dna ~naucki@236-233-dsl.kielnet.net
1156315896 J * Johnnie ~jdlewis@static-acs-24-154-32-33.zoominternet.net
1156317447 Q * cdrx Ping timeout: 480 seconds
1156317575 J * pisco ~pampel@p50879610.dip0.t-ipconnect.de
1156318342 Q * Aiken Ping timeout: 480 seconds
1156319379 J * cdrx ~legoater@242.32.96-84.rev.gaoland.net
1156320418 J * yarihm ~yarihm@whitehead2.nine.ch
1156321379 M * gcc__ I think i found a bug in the start-vservers script, anyone here knows it well?
1156321478 Q * bogus Quit: Reconnecting
1156321480 J * bogus ~bogusano@fengor.net
1156321489 M * harry gcc__: not me :)
1156321500 M * harry but i will look @ it :)
1156321506 M * gcc__ ok thanks
1156321524 M * gcc__ consider this code, if you will:
1156321538 M * gcc__ while test $i -gt 0; do
1156321538 M * gcc__     let --i
1156321538 M * gcc__     d=$__CONFDIR/${vservers[$i]}/apps/init
1156321538 M * gcc__     f=$d/mark
1156321538 M * gcc__     { test -n "$OPTION_MARK" -a -r "$f" && grep -qx "$OPTION_MARK" "$f"; } || \
1156321545 M * gcc__ { test -z "$OPTION_MARK" && test ! -e "$f"; } || \
1156321552 M * gcc__ unset vservers[$i]
1156321559 M * gcc__ done
1156321574 M * coocoon gcc__: why not using http://paste.linux-vserver.org/
1156321587 M * gcc__ ok, didn't know about that, i will
1156321596 M * harry coocoon: too hard :p
1156321626 M * gcc__ http://paste.linux-vserver.org/272
1156321628 M * coocoon harry: no really
1156321638 M * gcc__ this code is supposed to go through the list of vservers
1156321652 M * harry uhu\
1156321654 M * gcc__ and if the user specified a mark, like "default", remove all the ones that are not so marked
1156321667 M * gcc__ but, if no mark is specified, it ignores marked vservers
1156321687 M * gcc__ (removes them) and that means that start-vservers --all --stop does not stop marked vservers
1156321701 M * gcc__ because this loop removes them
1156321708 M * harry yes
1156321741 M * gcc__ I changed the "test -z" line to "{ test -z "$OPTION_MARK"; } || \" and now it works for me
1156321746 M * harry what's the problem with that?
1156321762 M * gcc__ start-vservers --all --stop is used at system shutdown, at least by my initscripts
1156321774 M * gcc__ it makes sense to me that it should stop ALL vservers, including marked ones
1156321776 M * gcc__ but it doesn't
1156321893 M * harry hmmm
1156321949 P * JimmyGulp Leaving
1156322251 M * Hollow gcc__: do you have http://daniel.hozac.com/vserver/util-vserver/util-vserver-0.30.210-start-vservers.patch aplied?
1156322330 M * gcc__ no, but it would appear to fix the problem
1156322335 M * gcc__ are there plans to merge it?
1156322336 M * Hollow yep ;)
1156322345 M * Hollow dunno.. enrico is not much around
1156322395 M * gcc__ are there a lot of important patches not applied to util-vserver?
1156322415 M * Hollow quite a few, yep
1156322419 M * gcc__ :-(
1156322423 M * Hollow you can find all in daniel hozacs space
1156322452 M * Hollow some are features, some fix bugs
1156322456 M * gcc__ gentoo applied a ton of them when i installed util-vserver
1156322475 M * Hollow yep, we have all of daniels patches in the ebuild
1156322481 M * harry it would be nice to have a list of all "important" patches to the utils... :S
1156322482 M * Hollow and he has all of ours :P
1156322500 M * harry or just a "good" version of the utils
1156322505 M * gcc__ hollow: then howcome i didn't get that patch?
1156322507 A * Hollow giggles
1156322518 M * harry where all the "old and deprecated" scripts are removed
1156322520 M * Hollow gcc__: which version do you have installed?
1156322568 M * gcc__ util-vserver-0.30.210-r13.ebuild
1156322587 M * Hollow heh.. the start-vservers patch was added in -r14
1156322599 M * gcc__ damn
1156322604 M * Hollow but we're at r17 currently
1156322610 M * gcc__ the only other one I got is r17, but it's masked for some reason
1156322636 M * Hollow yeah, because of the normal "wait wait some time before marking things as stable" guideline
1156322639 M * Hollow ;)
1156322644 M * gcc__ this is quite important though
1156322646 M * Hollow *we wait
1156322657 M * gcc__ i think it would be worth bumping the stable rev up to r14 at least
1156322684 M * Hollow i'll mark r17 stable, have to commit r18 anyway
1156322732 M * gcc__ btw i had to mess with my ebuild, to add --enable-apis=NOLEGACY, there's no use flag for that, right?
1156322791 M * Hollow hm, but it annoys me.. do you have legacy enabled or disabled in the kernel?
1156322810 M * Hollow s/hm/no/
1156322835 M * harry Hollow: so we'll get a new version of the tools when vs2.0.2 is released?
1156322850 M * gcc__ i have a very old kernel, over 1 year old
1156322871 M * Hollow harry: ask enrico ;)
1156322878 M * harry enricoooooooooooooooooooooo
1156322896 M * Hollow gcc__: ah, ok... that explains it...
1156322902 M * gcc__ ok gotta go, laters
1156322905 M * Hollow why don't use use recent kernels? ;)
1156322910 M * gcc__ that use flag would be really useful ;-)
1156322913 M * Hollow ah.. i can't type today
1156322934 M * gcc__ hollow, this box isn't easy to get physical access, and i don't want downtime
1156322944 M * Hollow ic.. will take a look
1156322945 M * gcc__ that's why i haven't upgraded for a while
1156322947 M * gcc__ i will soon
1156322954 M * gcc__ later all
1156322957 M * Hollow cu
1156322959 Q * gcc__ Quit: work
1156323058 M * harry a decent howto would also rule btw...
1156323075 M * harry i'm working with vservers for allmost a year now, and i still don't know if i'm doing stuff righ
1156323078 M * harry t
1156323081 M * harry :S
1156323093 M * sid3windr I think mine can be broken out of ;)
1156323100 M * sid3windr but ahwell. 
1156323107 M * Hollow harry: http://home.xnull.de/work/vserver/manual/manual.pdf
1156323108 M * Hollow ;)
1156323223 M * Hollow not much content yet though... (just two days old)
1156324158 Q * Greek0 Remote host closed the connection
1156324174 J * Greek0 ~greek0@85.255.145.201
1156325648 Q * shedi Quit: Leaving
1156325656 J * balbir ~balbir@59.145.136.1
1156326692 Q * zkbrsnie Read error: Connection reset by peer
1156327684 Q * yarihm iridium.oftc.net nobelium.oftc.net
1156327716 J * yarihm ~yarihm@whitehead2.nine.ch
1156327983 Q * glutoman Read error: Connection reset by peer
1156328008 J * glut glut@no.suid.pl
1156328103 J * shedi ~siggi@dsl-149-109-85.hive.is
1156328183 Q * michal_ Ping timeout: 480 seconds
1156328477 Q * cdrx Ping timeout: 480 seconds
1156329063 J * michal_ ~michal@www.rsbac.org
1156329371 J * fosco fosco@konoha.devnullteam.org
1156329373 M * fosco hi
1156329401 M * fosco is it possible to limit the memory of a vserver?
1156329408 J * lilalinux ~plasma@dslb-084-058-219-152.pools.arcor-ip.net
1156329797 M * fosco /etc/vservers/<vserver>/memlock does not seems to do anything
1156329798 M * coocoon fosco: yes it is have a look here  http://linux-vserver.org/Memory+Management http://linux-vserver.org/Memory+Allocation http://linux-vserver.org/some_hints_from_john
1156330172 M * fosco hmm thanks, seems http://linux-vserver.org/Memory+Allocation is obsolete
1156330212 M * coocoon fosco: u know openvcp
1156330224 M * Hollow why should it be obsolete?
1156330249 M * fosco because it talks about files that are not been used anymone
1156330270 M * fosco /etc/vservers/<vserver>/rss for example
1156330296 M * Hollow ah, right.. well, the page is about memory in general.. that section should just be fixed/updated
1156330299 M * fosco /etc/vservers/<vserver>/rlimits/rss for now
1156330340 M * doener I guess that was always broken, the paths listed there were never used IIRC
1156330347 M * doener fosco: feel free to fix it ;)
1156330368 M * fosco just a stupid question, if I want to limot, for example, 64MB of RSS, with a pagesize of 4K, what should I put un 'rss' ?
1156330372 M * fosco doener: why not :)
1156330401 M * Hollow 64*1024/4
1156330479 M * fosco ok thanks
1156330906 M * fosco (page updated)
1156330994 M * doener thanks :)
1156331150 J * cdrx ~legoater@cap31-3-82-227-199-249.fbx.proxad.net
1156333780 Q * pagano Ping timeout: 480 seconds
1156333947 J * pagano ~pagano@131.154.5.20
1156334310 M * daniel_hozac harry: http://daniel.hozac.com/vserver/util-vserver/ and http://archives.linux-vserver.org/200607/0139.html should tell you what patches to use.
1156334327 M * daniel_hozac (the mail is a bit old now though)
1156334357 M * derjohn daniel_hozac, I tried to apply you patch to the Debian sourcesm which are pre-patched with other stuff. unforunately they dont appy, even with fuzz 500 ...
1156334371 M * daniel_hozac derjohn: yeah, the configure hunk messes stuff up.
1156334375 M * daniel_hozac it doesn't even apply to my tree ;)
1156334392 M * daniel_hozac (which is why i'm seriously considering spinning a 0.30.210.1 while waiting for ensc...)
1156334464 M * daniel_hozac derjohn: filterdiff -x '*/configure' util-vserver-0.30.210-template.patch | patch -p1; autoconf might work better.
1156334491 M * daniel_hozac (if you have autoconf installed, that is)
1156334767 M * derjohn daniel_hozac, well, which autoconf version? I remeber it being picky about version 1.{4|5|6|7}
1156334776 M * daniel_hozac umm, the FC5 one :)
1156334785 M * derjohn daniel_hozac, then all is clear ;)
1156334791 M * daniel_hozac 2.59.
1156334808 M * derjohn ah uh oh --- i'll look and report. 
1156334831 M * daniel_hozac looks like anything >2.57 is ok though.
1156334835 M * derjohn wouldnt it make sense ensc gives you write access to cvs? maybe a -hozac branch?
1156334839 M * daniel_hozac according to configure.ac
1156334852 M * daniel_hozac i don't know if he's still using CVS.
1156334856 M * derjohn k , thx. will try to build a deb package
1156334858 M * daniel_hozac he hasn't touched the tree since 0.30.210.
1156334882 M * daniel_hozac (and i did get him to look at most of the patches back in july)
1156334966 M * daniel_hozac personally i think util-vserver should move to the linux-vserver.org infrastructure first though.
1156334969 M * derjohn or svn or whatever ... I just would like to see the 'current' branch in public read access and write access from you both to .. prevent a 'branch' to "util-vserver-ng" ..
1156334972 M * daniel_hozac i.e. svn.linux-vserver.org.
1156334977 M * derjohn full ack.
1156335013 Q * rob-84x^ Remote host closed the connection
1156335045 M * daniel_hozac would make a hozac branch much easier too.
1156335314 Q * yarihm Quit: Leaving
1156335485 J * crypt ~oli@195.35.72.197
1156335488 M * crypt Hi all
1156335511 M * daniel_hozac hello
1156335532 M * crypt are there some up to date patches for yum?
1156335548 J * rob-84x^ rob@submarine.ath.cx
1156335572 M * daniel_hozac which yum?
1156335598 M * daniel_hozac http://daniel.hozac.com/vserver/yum-2.6.1-chroot.patch
1156335632 M * crypt i'm searching this one and for yum 2.4
1156335656 M * daniel_hozac the yum 2.3.4 (IIRC) patch in contrib/ applies to 2.4.
1156335698 M * crypt ah ok thx a lot :)
1156335732 M * daniel_hozac ugh, autoconf... "configure.ac:27: version `0.30.210.1' doesn't follow Gnits standards
1156335841 M * doener wth is Gnits?
1156335852 M * daniel_hozac this is the first time i've ever heard of it.
1156336106 M * doener GNU nit-pickers gang... and they mess with "my" versioning scheme? oh well..
1156336127 Q * glut Read error: Connection reset by peer
1156336145 J * glut glut@no.suid.pl
1156336192 M * daniel_hozac i can't even find the gnits standards for version numbers...
1156336244 M * daniel_hozac but really, when it starts telling me what versions to use, it has seriously overstepped its bounds.
1156336770 Q * michal_ Ping timeout: 480 seconds
1156336884 Q * glut Ping timeout: 480 seconds
1156337120 J * glut glut@no.suid.pl
1156337152 J * michal_ ~michal@www.rsbac.org
1156337330 M * waldi hrm
1156337405 M * waldi is it possible that nfs interacts badly with vserver under low memory conditions? i have one buildd which regulary breaks the kernel and uses any data via nfs and another which works fine but uses local storage ...
1156337679 M * daniel_hozac does NFS and vanilla not show the problem?
1156337694 M * waldi nope
1156337732 M * waldi it does not show up with the non-vserver kernels
1156337735 M * daniel_hozac how does the kernel break?
1156337762 M * waldi many processes in D and Z state within the vserver
1156337817 M * daniel_hozac and nothing in dmesg?
1156337819 M * waldi hmm, no, the Z state process are generated by killing them
1156337835 M * waldi rm
1156337837 M * waldi Unable to handle kernel pointer dereference at virtual kernel address 0000000000000000
1156337844 M * waldi wtf
1156337880 M * daniel_hozac well, that explains it, i guess :)
1156337936 M * waldi hmm
1156338345 M * waldi the error is in generic code ...
1156338850 M * doener got a trace for us?
1156339031 M * waldi ([<000000000005bb04>] unqueue_me+0x38/0xdc) [<000000000005c4bc>] do_futex+0x2fc/0xbc8 [<000000000005d1f4>] compat_sys_futex+0x104/0x11c [<000000000001d864>] sysc_noemu+0x10/0x16 [<0000000040029bce>] 0x40029bce
1156339108 M * doener argh! *goes kicking php* (I'm debugging a php script that hangs on some futex for weeks now, the trace did jsut remind me ;)
1156339221 M * waldi hmm?
1156339246 M * doener second line is "do_futex"...
1156339250 Q * crypt Quit: Leaving
1156339319 M * doener waldi: fixed in mainline on Aug 5th
1156339337 M * waldi hmm, which version?
1156339404 M * doener must be rc4, rc3 was on July 29th
1156339431 M * doener (rc4 was Aug 6th)
1156339456 M * doener waldi: http://kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=e91467ecd1ef381377fd327c0ded922835ec52ab
1156339544 M * waldi ah, have to submit it for stable
1156339722 Q * balbir Ping timeout: 480 seconds
1156340311 J * balbir ~balbir@59.145.136.1
1156340340 N * Bertl_zZ Bertl_oO
1156341411 J * lilalinux_ ~plasma@h1-gw.of.net-lab.net
1156341777 Q * balbir Ping timeout: 480 seconds
1156341838 Q * lilalinux Ping timeout: 480 seconds
1156342189 J * mef ~mef@targe.CS.Princeton.EDU
1156342205 P * mef 
1156342557 J * marcfiu ~mef@targe.CS.Princeton.EDU
1156342817 Q * mire Ping timeout: 480 seconds
1156343141 J * balbir ~balbir@59.145.136.1
1156343318 Q * Johnnie Quit: G'bye!
1156343377 J * mire ~mire@144-167-222-85.COOL.ADSL.VLine.verat.net
1156343378 Q * balbir 
1156343407 J * Johnnie ~jdlewis@static-acs-24-154-32-33.zoominternet.net
1156344250 N * Belu_zZz Belu
1156345023 J * pisc1 ~pampel@p50879C63.dip0.t-ipconnect.de
1156345029 Q * meebey Ping timeout: 480 seconds
1156345302 Q * pisco Ping timeout: 480 seconds
1156345452 J * gerrit_ ~gerrit@bi01p1.co.us.ibm.com
1156345551 J * meebey meebey@booster.qnetp.net
1156346054 Q * olilo Remote host closed the connection
1156346102 J * olilo hiddenserv@tor.noreply.org
1156346393 Q * insomniac Read error: Connection reset by peer
1156347227 J * stefani ~stefani@tsipoor.banerian.org
1156347289 J * lolilol hiddenserv@tor.noreply.org
1156347364 Q * olilo Ping timeout: 480 seconds
1156348344 Q * marcfiu Quit: Download Gaim: http://gaim.sourceforge.net/
1156349951 J * bonbons ~bonbons@83.222.36.236
1156351316 J * mef ~mef@c-68-39-177-97.hsd1.nj.comcast.net
1156351322 P * mef 
1156352079 J * marcfiu ~mef@c-68-39-177-97.hsd1.nj.comcast.net
1156352770 Q * shedi Quit: Leaving
1156354095 M * cryptronic hi all
1156354098 M * cryptronic http://paste.linux-vserver.org/273
1156354116 M * cryptronic i searched for help, and found that i can reduce the timeout
1156354129 M * cryptronic but where do i have to reduce the timeout?
1156354161 N * Bertl_oO Bertl
1156354168 M * Bertl evening folks! 
1156354185 M * doener hi Bertl 
1156354189 M * Johnnie Bertl ;)
1156354198 M * cryptronic hi bertl
1156354235 M * cryptronic Bertl, i got following error: http://paste.linux-vserver.org/273 and you wrote in a ml, that i can reduce the timeout, but where do i have to reduce this timeout?
1156354258 M * Bertl cryptronic: thing is, you probably want to fix the cause, not treat the symptoms :)
1156354270 M * Bertl hey Johnnie! LTNS!
1156354283 M * cryptronic ok ;) but i have no clue where the errors came from :(
1156354329 M * Bertl cryptronic: me neither yet, it looks like your stop script itself gets killed by the timeout
1156354341 M * cryptronic ok
1156354347 M * cryptronic that's bad
1156354370 M * Bertl first, what kernel version and tools? (best testme.sh uploaded to paste.linux-vserver.org)
1156354382 M * cryptronic just a moment
1156354396 M * cryptronic btw: whole stop failing: http://paste.linux-vserver.org/274
1156354436 M * cryptronic testme.sh output: http://paste.linux-vserver.org/275
1156354454 M * cryptronic i use the kernel from derjohn provided on his homepage
1156354472 M * Bertl okay, looks fine so far ...
1156354519 M * Bertl I would supsect your shutdown scripts (probably rc + friends?) which try a few things they are not allowed to do (as you see from the shutdown), finally call reboot -f
1156354550 M * cryptronic these things are set in the vserver right?
1156354551 M * Bertl I further assume that your guest has somehow reboot_kill enabled (could you verify that?)
1156354568 M * Bertl daniel_hozac: is reboot_kill default now?
1156354585 M * Bertl (or did we make it default for certain guests :)
1156354713 M * cryptronic wheer do i find this setting?
1156354732 J * shedi ~siggi@inferno.lhi.is
1156354784 M * Bertl it's part of the flags, but it might be that we set them by default (depending on the guest creation) best check on a running guest with /proc/virtual/*
1156354804 J * Viper0482 ~Viper0482@p549778C0.dip.t-dialin.net
1156354939 M * cryptronic http://paste.linux-vserver.org/276
1156354957 M * Bertl *phone* sec
1156355170 Q * s0undt3ch Quit: leaving
1156355176 J * s0undt3ch hbsiqimy@bl7-241-134.dsl.telepac.pt
1156355664 Q * marcfiu Ping timeout: 480 seconds
1156355813 N * abi` abi
1156356116 J * zkbrsnie ~zkbrsnie@83-64-146-226.klosterneuburg.xdsl-line.inode.at
1156356573 Q * Nam Ping timeout: 480 seconds
1156356648 M * daniel_hozac Bertl: i can't see where that would be.
1156359051 M * cryptronic Bertl, how long are you phoning ;)
1156359072 M * phedny my personal phone call record is just below 2 hours :)
1156359114 M * cryptronic ah ;)
1156359605 M * Hollow daniel_hozac: what was the fix for "No command given" again?
1156359996 M * cehteh # vserver etch_test start
1156359996 M * cehteh vshelper.init: can not determine xid of vserver 'etch_test'; returned value was ''
1156360001 M * daniel_hozac Hollow: hmm?
1156360009 M * cehteh mhm .. what can be the cause of that?
1156360023 M * daniel_hozac cehteh: your guest doesn't start any services and it's not using a plain initstyle.
1156360027 J * EleRas ~EleRas@p508497A1.dip0.t-ipconnect.de
1156360041 M * cehteh ah maybe
1156360056 M * Hollow daniel_hozac: the No command given on vserver .. start
1156360062 M * EleRas hi
1156360064 M * Hollow i saw this problem quite a few times now
1156360068 M * Hollow wasn't there some fix for it?
1156360073 M * daniel_hozac Hollow: sysv initstyle and no /etc/{rc,init}.d/rc?
1156360085 M * Hollow ah.. wrong init style..
1156360089 M * Hollow yep, that my be it
1156360093 M * Hollow EleRas: which init style do you have?
1156360102 M * EleRas oh, good question
1156360115 M * EleRas i think it's the default
1156360120 M * EleRas sysv
1156360129 M * Hollow isn't plain the default?
1156360135 M * daniel_hozac no.
1156360137 M * Hollow oh
1156360143 M * Hollow yeah, then you probably need plain init style
1156360160 M * EleRas ok, i'll try this, thx
1156360184 M * Hollow yw!
1156360201 M * cehteh daniel_hozac: thanks .. that was it, quite confusing thing
1156360266 M * daniel_hozac cehteh: yeah, that error needs to be improved...
1156360271 J * comfrey ~comfrey@h-64-105-215-75.sttnwaho.covad.net
1156360306 M * cehteh the error message, not the error ;)
1156360322 M * daniel_hozac lol, right.
1156360328 M * Hollow just like the no command one ;)
1156360367 M * Hollow daniel_hozac: btw, with your latest patches (clone + template) our vserver-new script will be obsolete, right?
1156360377 M * Hollow (except the syslog convenience)
1156360384 M * daniel_hozac right.
1156360384 M * Hollow in stage3 target
1156360391 M * cehteh soo ... building a via C3 optimized glibc ... 3rd try ...
1156360400 M * daniel_hozac but -m clone is pretty hacky.
1156360420 M * Hollow ok... phreak``: we should probably remove the vserver-new script, or at least bail out and refer to vserver build -m template
1156360424 M * cehteh (learning how to build .deb's by try'n'error and starting with glibc is prolly the hard way)
1156360440 M * Hollow daniel_hozac: hacky in what sense?
1156360524 M * daniel_hozac in the sense that i think it's likely very easy to break it.
1156360563 M * Hollow well, i guess that was the case with vserver-new too.. just a simple rsync in this case iirc
1156360628 M * daniel_hozac IMHO clone really should get some helper to only copy files that aren't iunlink|immutable and just link those that are.
1156360650 M * Hollow basically a CoW aware clone
1156360659 M * daniel_hozac not necessarily COW.
1156360664 M * daniel_hozac it would be useful even on stable.
1156360950 M * Bertl *phew* longer call :)
1156360968 M * phedny almost 2 hours ;)
1156360976 M * cryptronic Bertl, no problem
1156360987 M * cryptronic hopefully you could help me
1156360999 M * Bertl yeah, and oversee, glad that I have not to pay for it :)
1156361008 M * cryptronic ;)
1156361062 M * Bertl phedny: so I take it, I did beat your record there :)
1156361079 M * phedny don't know exact number of minutes, but it could be
1156361094 M * cryptronic ;)
1156361100 M * phedny 2 only needed two ingredients for my record: a girlfriend and free voip calls :)
1156361203 M * Bertl reading up now ...
1156361231 M * cryptronic thx, if you mean my paste ;)
1156361405 M * EleRas thx Hollow & daniel_hozac, the initstyle = plain did the trick
1156361870 M * daniel_hozac Bertl: what happened to 2.0.2? :)
1156361903 A * mnemoc raise his ears
1156361932 M * Bertl daniel_hozac: well, I had no time to figure the changes to 2.0.1 yet ...
1156361941 M * daniel_hozac all of them, or what?
1156361949 M * Bertl daniel_hozac: but you can consider it released :)
1156361965 M * daniel_hozac http://daniel.hozac.com/vserver/2.0.2.deltas has deltas from -rc1+.
1156361970 M * cryptronic Bertl, could you take maybe a look on my problem again?
1156361980 M * Bertl cryptronic: I'm doing that right now
1156361988 M * cryptronic thx :)
1156361996 M * daniel_hozac cryptronic: what distro is that?
1156362026 M * cryptronic hostsystem: debian unstable vserver: ubuntu dapper
1156362099 M * daniel_hozac cryptronic: are you sure your guest's last initscript doesn't do killall5 -9?
1156362131 M * cryptronic just a second
1156362233 M * Bertl yes, it's probably the sysv shutdown which 'kills' itself
1156362253 M * cryptronic hmm where do i find that entry?
1156362492 M * Bertl check what scripts are executed, maybe it's the rc script itels
1156362494 M * Bertl *itself
1156362499 M * cryptronic ok
1156362506 M * Bertl but more likely it's the 'halt' script
1156362609 M * cryptronic in the init script itself it says only reboot but no killall
1156362693 M * Bertl reboot_kill is disabled, so there should be no killing
1156362730 M * Bertl daniel_hozac: ad changes, cool, tx!
1156362760 M * cryptronic hmm
1156362814 M * cryptronic when the vserver shut's down it goes into rc6 which ends with reboot
1156362850 M * Bertl cryptronic: could you upload the output of:
1156362878 M * Bertl  ls -la /etc/rc.d/rc6.d/
1156362882 M * Bertl (inside the guest)
1156362933 M * cryptronic http://paste.linux-vserver.org/277
1156362982 M * Bertl okay, now let's remove the following entries (links) there
1156363019 M * Bertl S30urandom, S31umountnfs.sh, S35networking, S40umountfs, S60umountroot
1156363038 M * Bertl and upload the contents of /etc//init.d/sendsigs
1156363080 M * Bertl daniel_hozac: maybe we should protect the shutdown script somehow?
1156363099 M * Bertl daniel_hozac: btw, did you see the spectator patch?
1156363153 M * daniel_hozac hmm, no.
1156363179 M * daniel_hozac how would we identify the shutdown script?
1156363226 M * cryptronic http://paste.linux-vserver.org/278
1156363270 M * Bertl here we have the killall5 -9
1156363338 M * Bertl try to remove that link too, and see how itgoes
1156363440 M * cryptronic ok the error went away after deleting the symlinks you mentioned
1156363454 J * yarihm ~yarihm@84-75-121-93.dclient.hispeed.ch
1156363466 M * cryptronic but when i now have multiple ips only the first ip gets turned off
1156363481 M * Bertl known mainline issue
1156363492 M * cryptronic ah ok
1156363502 M * Bertl hmm, wait, what did you mean?
1156363552 M * Bertl (because I think I interpreted something into your line which is not true :)
1156363573 M * cryptronic when i have interface 0 and 1. i start the vserver and the 2 ips are up, after stopping the vserver only the ip of 0 is disabled ip of 1 is still up
1156363603 M * Bertl okay, in this case, check that '1' does not contain a nodev line
1156363609 M * Bertl s/line/entry
1156363627 M * Bertl (i.e. it has a dev entry)
1156363639 Q * lilalinux_ Remote host closed the connection
1156363641 M * cryptronic ok when i include a nodev line ifconfig in the guest only contaibns a eth0 with the ip of the last nodev
1156363651 J * DreamerC_ ~dreamerc@59-112-1-201.dynamic.hinet.net
1156363661 M * Bertl nodev means that the tools do not add/remove the ip for you
1156363689 M * Bertl so, if you want the tools to handle the ips, you have to give a dev entry
1156363706 M * cryptronic no nodev does not exist
1156363712 M * Bertl and then you should make sure that the ip/prefix/mask matches the currently configured
1156363733 M * Bertl (which will be true when the guest created that ip for you)
1156363734 M * cryptronic for 1 i have set name = 1
1156363743 M * Bertl that's fine
1156363744 M * cryptronic yes i have set that all
1156363760 M * Bertl so tools will create _and_ remove the ips propery
1156363763 M * Bertl *properly
1156363771 M * cryptronic :(
1156363794 M * cryptronic when a ip already is up it won't get managed or?
1156363795 M * Bertl I'm pretty sure your current ip (which seems to be left) does not match the configured
1156363808 M * Bertl i.e. different netmask/prefix
1156363830 M * Bertl you probably also get an error message on startup
1156363850 M * cryptronic RTNETLINK answers: File exists
1156363855 M * Bertl precisely
1156363864 M * cryptronic how can i remove that?
1156363873 M * Bertl remove the ip manually 
1156363890 M * Bertl on next startup it will get the 'right'(the configured) ip/mask
1156363896 M * cryptronic just a moment
1156363956 J * Nam ~nam@70.71.224.66
1156363963 M * Bertl wb Nam!
1156363965 M * cryptronic how can i remove that ip if it doesn't show up in ifconfig in the host?
1156363974 M * Bertl with ip?
1156363975 M * michal_ ip addr list and than
1156363988 M * Bertl ifconfig is a legacy tool
1156363988 M * michal_ ip addr del <ipaddr>/<netmask>
1156363992 M * cryptronic ahhh :)
1156363994 M * cryptronic thx a lot
1156364025 M * michal_ ...and netlink is a beafutiful interface :]
1156364036 M * michal_ *beautiful
1156364050 M * cryptronic now i know what we did wrong in openvcp
1156364052 Q * DreamerC Ping timeout: 480 seconds
1156364104 M * cryptronic thx a lot for your help Bertl 
1156364108 M * Bertl np
1156364110 M * michal_ anyway i should behave.
1156364123 M * michal_ hello Bertl, hello #vserver :]
1156364133 M * Bertl hey michal_!
1156364235 J * chris___ ~chris@cpc1-cmbg7-0-0-cust497.cmbg.cable.ntl.com
1156364275 M * chris___ hi all, does ps on the host normally show processes in vservers? if not, is there a hacked ps that does
1156364306 M * Bertl chris___: no and no, but there is vps
1156364341 Q * pisc1 Quit: Download Gaim: http://gaim.sourceforge.net/
1156364400 M * chris___ ok, thanks. i have a ton of processes in one vserver, is there some way to limit the number of processes? sorry if this is a FAQ.
1156364438 M * Bertl http://linux-vserver.org/Resource+Limits
1156364447 M * chris___ thanks
1156364452 M * Bertl nproc is what you want here
1156364471 M * Bertl you can apply that while it is running via vlimit
1156364523 M * chris___ can i make such limits apply by default when the vserver is started?
1156364577 Q * ntrs__ Read error: Connection reset by peer
1156364578 J * ntrs__ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com
1156364634 M * Bertl daniel_hozac: http://paste.linux-vserver.org/279
1156364674 M * Bertl chris___: yep, you add them to the config: here are the details: http://www.nongnu.org/util-vserver/doc/conf/configuration.html
1156364698 Q * bonbons Quit: Leaving
1156364713 M * chris___ thanks! this vserver has gone awol, any way to kill the excess processes without restarting it?
1156364726 M * Bertl vkill is your friend :)
1156364755 M * chris___ Bertl, that web page has the most horrible color scheme that I have ever seen
1156364762 M * chris___ it's worse than the rasta kernel debugger
1156364768 M * Bertl daniel_hozac: we need to get some details on what exactly was fixed ...
1156364778 M * Bertl chris___: the stylesheet I use looks quite nice
1156364788 M * cehteh haha
1156364819 M * cehteh can you safe the stylesheel preference in firefox for a page?
1156364835 M * Bertl don't know, my galeon always remebers
1156364868 A * cehteh should switch back to galeon ...
1156364881 M * chris___ Bertl, what do i have to insert, and into which file, to limit nproc on a vserver? i can't understand that page.
1156364898 M * Bertl well, we are talking about limits, right?
1156364918 M * cehteh i only switched from sylpheed to thunderbird and from galeon to ff for knowing them that i can help my wife with them
1156364928 M * chris___ something under /etc/vservers/<name>/rlimits/nproc.hard?
1156364963 M * Bertl there you go
1156364973 M * chris___ just echo e.g. 50 into that file?
1156364982 M * Bertl <name> is the guest name and whatever you echo there is the limit
1156365016 M * Bertl you can check that after guest restart in /proc/virtual/<xid>/limit
1156365554 M * chris___ do vservers not normally run the sysv startup scripts, e.g. /etc/rcS.d on debian?
1156365708 M * Bertl depends on the setup, usually they run the runlevel scripts (with sysv init style) or 'just' start init (with the plain init style)
1156365735 M * Bertl derjohn: ping?
1156365736 M * chris___ if u don't specify when running vserver build, what do you get by default, and how do you change it?
1156365752 M * Bertl you get the plain init style by default, IIRC
1156365773 M * Bertl and you can change it in the config with echo "sysv" > ...
1156365812 M * Bertl /etc/vservers/<name>/apps/init/style
1156365890 M * chris___ great, thanks!
1156365953 M * Bertl you're welcome!
1156365962 M * chris___ hmm, still doesn't run /etc/init.d/hostname, as far as I can tell
1156365993 M * chris___ which is linked from /etc/rcS.d, not /etc/rc3.d
1156366036 M * Bertl well, it probably doesn't execute the rcS runlevel unless explicitely specified
1156366054 M * Bertl as for a guest the system is already running
1156366060 M * chris___ how would I do that?
1156366097 M * Bertl you can either specify a different runlevel in the config, or cross link them, or call one from the other
1156366242 M * chris___ ok, will do, thanks
1156366266 M * chris___ another question, sorry: when I shut down this debian vserver, I get the following error:
1156366267 M * chris___ /usr/lib/util-vserver/vserver.stop: line 85:  7498 Killed                  "${NICE_CMD[@]}" ${USE_VNAMESPACE:+$_VNAMESPACE --enter "$S_CONTEXT" -- } $_VCONTEXT $SILENT_OPT --migrate --chroot --xid "$S_CONTEXT" -- "${INITCMD_STOP[@]}"
1156366271 M * Bertl btw, you can set the hostname from outside too
1156366272 M * chris___ any ideas?
1156366293 M * Bertl please read up on the irc log, we handled that a few minutes ago
1156366321 Q * DreamerC_ Quit: leaving
1156366322 M * Bertl it boils down to removing the killall5 from the shutdown
1156366332 M * chris___ sorry, i wasn't here at the time, is it online somewhere?
1156366337 J * DreamerC ~dreamerc@59-112-1-201.dynamic.hinet.net
1156366345 M * Bertl (or to remove the rc6.d/* links which are hardware related
1156366365 Q * Viper0482 Remote host closed the connection
1156366366 M * Bertl http://irc.13thfloor.at/LOG/2006-08/LOG_2006-08-23.txt
1156366383 M * cryptronic Bertl, i'll write a hwoto about that
1156366397 M * Bertl even better! :)
1156366405 M * cryptronic ;)
1156366412 M * chris___ ok, removing S90reboot fixed that error, thanks
1156366418 M * Bertl in general, when you use sysv style, you want to remove _everything_ hardware related
1156366432 M * Bertl i.e. only services should be listed
1156366435 M * chris___ i still get a cosmetic error that a timeout occurred while shutting down, but no processes are listed as running
1156366461 M * Bertl that's at least interesting (that no process is shown)
1156366470 M * chris___ and start-stop-daemon doesn't work either (I guess debian is badly broken in vservers)
1156366471 M * cryptronic Bertl, you helped me a lot and it is very helpful to have such howtos ;) so i'll write that down tomorrow
1156366494 M * Bertl chris___: it works quite fine with recent packages, what packages are you using?
1156366509 M * chris___ debian packages? or util-vserver?
1156366518 M * Bertl debian vserver packages :)
1156366519 M * cryptronic chris___, try out using openvcp (www.openvcp.org) we'll provide some images soon
1156366526 M * chris___ this system was installed for UML, i'm converting my VMs to vserver one at a time
1156366539 M * cryptronic ah ok
1156366565 M * Bertl chris___: what was the reason for switching from UML to Linux-VServer if I may ask?
1156366575 M * chris___ system load and performance
1156366609 M * chris___ i'm running about 8 vms on one machine, with UML the host (and all VMs) would become unusable when I ran my nightly backups
1156366614 M * chris___ and at random times during the day
1156366638 M * Bertl what hardware?
1156366645 M * chris___ amd64, 1 gig ram
1156366657 M * chris___ 3200+
1156366671 M * Bertl okay, so probably the memory was the main issue
1156366697 M * chris___ could well be, I/O wait was over 90% at such times
1156366704 M * chris___ even though the system had no swap configured
1156366813 M * Bertl well, you probably can run 20 cpu bound guests on that hardware easily
1156366854 M * chris___ with vservers or UML?
1156366863 M * Bertl with Linux-VServer :)
1156366870 M * chris___ i hope so :-)
1156366878 M * chris___ at least i hope my vservers will be mainly idle
1156366983 Q * meandtheshell Quit: bye bye ... 
1156366998 M * Bertl so back to your distro, what packages do you use there?
1156367002 M * Bertl (on the host)
1156367022 M * chris___ this is a gentoo host, i'm not even sure how to list all packages :-(
1156367046 M * Bertl ah, gentoo, so why do you complain about bad debian support then?
1156367068 M * chris___ debian guest
1156367079 M * chris___ i have a lot of diverse guests
1156367083 M * cryptronic i have running debian guest on my gentoo machine very well
1156367093 M * Bertl well, it's a guest, it does not support/know about the isolation
1156367104 M * chris___ cryptronic, did u install using vserver build -m debootstrap?
1156367118 M * cryptronic yes
1156367135 M * chris___ ok, that's probably why u got the right packages and i didn't :-)
1156367163 M * cryptronic ;)
1156367175 M * Bertl chris___: well, as far as I understood, the guest was already running with UML, no?
1156367187 Q * gerrit_ Ping timeout: 480 seconds
1156367206 M * chris___ yeah
1156367211 M * chris___ and working pretty well
1156367219 M * chris___ uml provides hardware virtualisation
1156367231 M * Bertl chris___: so if you (for a start) ignore the error messages it thows, it should run out of the box with 'plain' init style
1156367238 M * Bertl *throws
1156367259 M * chris___ yeah, it does boot, modulo the wrong hostname
1156367269 M * chris___ i'm trying to optimise it and clean up the errors it throws
1156367277 M * chris___ so i can tell what's really wrong, if anything
1156367296 M * Bertl that's step two, but that does not affect the functionality
1156367331 M * Bertl the hostname should be correct with 'plain' initstlye too, because the guest should know what to set it to, no?
1156367372 M * Bertl note: I'm not suggesting to go for the 'plain' init at all, for me that is just the 'straight forward' approach with all it disavantages
1156367392 M * Bertl a cleaned up sysv startup is the better way IMHO
1156367408 M * chris___ i'm working on it
1156367419 M * chris___ what are the differences between init styles?
1156367427 M * cryptronic chris___, : http://www.cryptronic.de/wiki/Vserver_en:images_for_openvcp here is a ready image you onmly have to remove the symlinks in rc6
1156367437 M * cryptronic you can use them as normal images
1156367439 M * Bertl the 'plain' just calls 'init' inside the guest, the rest is done by the guest
1156367456 M * Bertl the 'sysv' executes the runlevel scripts, no init is spawned
1156367517 A * Belu is away (i�ll be back later...)
1156367517 N * Belu Belu_zZz
1156367560 M * cryptronic i'll be off
1156367562 M * cryptronic cya
1156367569 M * Bertl cya!
1156367600 M * cryptronic thx again Bertl 
1156367616 M * Bertl np
1156367765 J * ntrs_ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com
1156367766 Q * ntrs__ Read error: Connection reset by peer
1156367783 Q * dna Quit: Verlassend
1156368055 M * daniel_hozac Bertl: did you see the bug on the ML?
1156368075 M * Bertl no, which one?
1156368086 P * EleRas Time makes no sense
1156368087 J * aminorex ~aminorex@71-13-40-131.dhcp.dlth.mn.charter.com
1156368090 M * daniel_hozac http://archives.linux-vserver.org/200608/0101.html
1156368103 M * daniel_hozac (i haven't gotten the email yet...)
1156368131 M * daniel_hozac i guess we'll want to fix that for 2.0.2.
1156368177 M * Bertl well, I haven't gotten the mail either, maybe we should talk to Martin, IMHO the ML has some issues
1156368280 M * Bertl shouldn't be relevant for 2.0.2, no?
1156368312 M * daniel_hozac how so?
1156368444 Q * Nam Ping timeout: 480 seconds
1156368465 M * Bertl ah, we already have the 'new' interfaces there
1156368485 M * Bertl somehow I was attributing that to 2.1.x
1156368491 M * daniel_hozac haven't they been around since 1.9.x? or maybe even 1.3.x.
1156368509 M * Bertl could be, I'm living behind the moon it seems :)
1156368528 M * michal_ bertl is living pink floyd style
1156368548 M * michal_ "...i meet you on the dark side of the moon"
1156368549 M * michal_ :P
1156368581 M * Bertl daniel_hozac: yeah, I guess that qualifies as bug fix, so it will get in
1156368608 M * Bertl daniel_hozac: nevertheless I'm somewhat worried about the mailing list ...
1156368612 Q * yarihm Quit: Leaving
1156368668 J * yarihm ~yarihm@84-75-121-93.dclient.hispeed.ch
1156368807 M * Bertl I'm missing some of the pre-start/stop mails too, and the begin of the IP alias limitation thread
1156368923 M * Bertl ah, well, martins host got listed on a spam list again :(
1156368946 M * Bertl I guess I'll add an override for that one for now ...
1156368964 M * daniel_hozac yeah, i keep missing emails as well.
1156368986 M * daniel_hozac i usually get them after a day or so though.
1156369121 Q * yarihm Quit: Leaving
1156369200 M * chris___ does anyone have any ideas why tomcat won't start automatically in a vserver, but when I enter it and run /etc/init.d/tomcat start it works fine?
1156369212 M * daniel_hozac bad PATH?
1156369219 M * chris___ it says it's starting during vserver start, but doesn't
1156369220 M * Bertl yes, that is a problem we still see
1156369234 M * Bertl debian guests still suffer from that
1156369257 M * chris___ i think it executes java, i changed the command to strace -f tomcat and it hangs on startup, vserver enter shows java hung and the strace still running
1156369265 M * chris___ this is actually an FC2 guest
1156369303 M * Bertl that's unusal, on debian guest I fixed that with:
1156369334 M * daniel_hozac may i ask what makes you use an EOL distribution?
1156369345 M * chris___ daniel_hozac, i haven't upgraded it yet
1156369350 M * Bertl http://paste.linux-vserver.org/280
1156369364 M * chris___ i trust FC2 a lot, it's worked very well
1156369390 M * Bertl daniel_hozac: could we get similar as 'sane' default for the sysv method into the tools?
1156369395 M * daniel_hozac but FC2 is completely dead now.
1156369401 M * chris___ daniel_hozac, as of about 2 weeks ago
1156369412 M * chris___ i didn't realise that fedora legacy would pull support like that
1156369413 M * daniel_hozac chris___: right.
1156369420 M * chris___ now i'm looking to move away from fedora completely
1156369433 M * daniel_hozac chris___: no interest -> no support.
1156369435 M * chris___ but upgrading all your machines to a different distro is not a trivial task
1156369456 M * daniel_hozac Bertl: i guess it should be possible.
1156369459 M * Bertl chris___: Linux-VServer helps here a lot
1156369467 M * chris___ i find it hard to believe that every single fedora user is quite comfortable with upgrading every 6 months
1156369480 M * Bertl chris___: you make a copy of the guest and try to simply 'upgrade' the guest from inside
1156369490 M * Bertl chris___: if that fails, no harm done :)
1156369510 M * daniel_hozac chris___: i tend to skip a release.
1156369536 M * daniel_hozac so it becomes more like 12-15 months.
1156369546 M * chris___ is there a fedora that's known to work well with java? or should i move to centos 4.3? or something else?
1156369555 M * chris___ daniel_hozac, still way too fast for me
1156369561 M * daniel_hozac Java sucks. nothing works well with it...
1156369577 M * chris___ i want to keep using an OS until i get annoyed with how outdated it is
1156369603 M * daniel_hozac i agree, it is quite fast.
1156369603 M * chris___ not be forced by lack of security updates
1156369618 M * daniel_hozac chris___: Legacy is always looking for more man-power...
1156369651 M * chris___ maybe i would contribute if i wasn't so annoyed with fedora's fast release cycle and how broken new releases are
1156369711 M * Bertl I can assure you that other distros (here Mandriva) are getting more and more broken too
1156369750 M * chris___ i don't like mandriva either, since mandrake decided you couldn't have the devel packages unless you pay for them
1156369754 M * Bertl the only distro which seems to really improve (regarding brokenness) is debian (only my limited view/opinion)
1156369783 M * chris___ Bertl, i'm beginning to agree that debian and ubuntu are the most viable distros for long-term support
1156369798 M * chris___ although debian annoys me with the weirdness of its init system
1156369804 M * Bertl chris___: Mandriva Community is all available, no fee required
1156369807 M * daniel_hozac CentOS is 5 years, right?
1156369828 M * chris___ yeah, centos is looking like a good option as well
1156369833 M * Bertl chris___: have you tried gentoo yet? (regarding init system weirdness :)
1156369866 M * chris___ well, at least gentoo gives you advantages for uits weirdness
1156369872 M * chris___ like super fast boot
1156369900 M * chris___ and /sbin/runscript is actually kind of cool, much more logical than plain sysv
1156369908 M * chris___ and i like rc-update syntax
1156369926 M * chris___ but gentoo keeps changing under my feet, and emerge sync takes bloody forever
1156370000 M * daniel_hozac didn't they improve that recently?
1156370004 M * daniel_hozac (emerge sync)
1156370044 M * chris___ not for me
1156370052 M * chris___ took about 90 mins the other day
1156370059 M * daniel_hozac hehe.
1156370093 M * chris___ Bertl, just tried your hack on /etc/rc on this machine, java still won't start at boot
1156370100 M * Bertl daniel_hozac: we didn't have lock isolation in 2.0.1, did we?
1156370107 M * chris___ the weird thing is that it does write to /home/tomcat/logs/catalina.out
1156370109 M * daniel_hozac /proc/locks?
1156370116 M * chris___ so it must be dying somewhere
1156370131 M * Bertl daniel_hozac: the entire locks/xid stuff
1156370158 M * daniel_hozac we did.
1156370161 M * Bertl ah, nah, it's there
1156370185 M * chris___ is that important for java? strace showed it hung in futex(...)
1156370275 M * Bertl are you sure that the dependancies are met, i.e. maybe you start something after that, and on enter/restart that is already running? (java sucks in so many regards)
1156370286 M * chris___ quite possibly
1156370318 M * chris___ but the only things that /usr/sbin/vserver start shows starting after tomcat are crond, mysql and smbus
1156370332 M * Bertl mysql?
1156370351 M * daniel_hozac Bertl: i guess we should mention the ccaps fix.
1156370352 M * chris___ shouldn't be required to start java, even though some VMs may need it
1156370410 M * Bertl daniel_hozac: http://paste.linux-vserver.org/281
1156370423 M * chris___ http://paste.linux-vserver.org/282
1156370426 M * chris___ any ideas?
1156370465 M * Bertl segfault? sounds like a java bug :)
1156370468 M * chris___ maybe i'll try upgrading java
1156370496 M * chris___ need i say that it works fine in uml? :-)
1156370535 M * Bertl with the same kernel version?
1156370553 M * Bertl futex changed a _lot_ over the years :)
1156370558 M * chris___ no, completely diff kernel
1156370573 M * chris___ that was a 2.4 UML VM, the host is now 2.6
1156370638 M * Bertl so you might want to compare it to 2.4.32 vserver then :)
1156370659 M * chris___ unfortunately i can't :-)
1156370669 M * chris___ ok this is weird
1156370672 M * daniel_hozac 2.4.33.1 is released now, btw ;)
1156370684 M * Bertl yeah, willy is doing a goodjob :)
1156370696 M * chris___ trying to wget a new java in the vm: Cannot write to `jdk-1_5_0_08-linux-i586-rpm.bin... (No space left on device).
1156370707 M * Bertl dlimits?
1156370714 M * daniel_hozac or really out of space.
1156370717 M * daniel_hozac such as on /tmp.
1156370734 M * Bertl who would get large packets to /tmp :)
1156370740 A * michal_ 
1156370742 M * Bertl *packages
1156370742 M * michal_ ;p
1156370748 M * michal_ i have it on tmpfs
1156370753 M * chris___ ok tmpfs, my bas
1156370758 M * chris___ s/bas/bad/
1156370775 M * chris___ didn't realise it would mount tmpfs, it didn't under uml
1156370779 M * michal_ mount /tmp -o remount,size=newsize ;)
1156370808 M * chris___ root@qwirx(/)# mount /tmp -o remount,size=128M
1156370808 M * chris___ mount: permission denied
1156370810 M * daniel_hozac tmpfs on /tmp is a vserver thing.
1156370816 M * chris___ i can't umount it either
1156370844 M * michal_ i think you should do it from outside but in a proper namespace
1156370848 M * daniel_hozac yep.
1156370851 M * mnemoc does 'secure_remount' let you do that?
1156370858 M * daniel_hozac hopefully not...
1156370864 M * Bertl if you really want a larger tmpfs (or none at all), change that in the fstab
1156370876 M * Bertl (of the configuration, that is)
1156370891 Q * mire Quit: Leaving
1156370915 M * daniel_hozac mnemoc: IMO it does.
1156370947 M * Bertl daniel_hozac: reiserfs dlimits? (did we have that on 2.01)?
1156370967 M * daniel_hozac yep.
1156371070 J * gerrit_ ~gerrit@c-67-160-146-170.hsd1.or.comcast.net
1156371094 M * chris___ still segfaults under strace with latest java
1156371118 M * Bertl latest java means?
1156371120 M * daniel_hozac chris___: tried using gij?
1156371136 M * chris___ jdk1.5.0_08
1156371151 M * chris___ gij causes weird problems with tomcat last time i checked
1156371155 M * chris___ maybe they fixed it by now
1156371156 M * Bertl chris___: for a quick test, you could fake the kernel version (and make it 2.4 something)
1156371178 M * chris___ yeah it could be NPTL
1156371200 M * chris___ LD_ASSUME_KERNEL=2.4?
1156371219 M * daniel_hozac i guess Bertl meant utsname faking.
1156371230 M * chris___ dunno how to do that, sorry
1156371235 M * Bertl utsname/uname --kernel-version
1156371248 M * Bertl --kernel-release
1156371261 M * Bertl you can adjust them in the config quite easily (echo)
1156371287 P * chris___ Leaving
1156371295 J * chris___ ~chris@cpc1-cmbg7-0-0-cust497.cmbg.cable.ntl.com
1156371321 M * chris___ sorry, still no idea, i don't have any dir called utsname, only utsname.h headers
1156371330 M * daniel_hozac mnemoc: it works here, i.e. i can make the filesystem however large i want ;)
1156371354 M * mnemoc daniel_hozac: no limits? ,-)
1156371394 M * daniel_hozac chris___: echo 2.4.32 > /etc/vservers/.../uts/release
1156371413 M * daniel_hozac mnemoc: memory limits? no.
1156371429 M * chris___ daniel_hozac, can i have that done automatically during vserver boot?
1156371449 M * daniel_hozac chris___: what? you set it in the configuration and the tools will do it for you, yes.
1156371453 M * mnemoc daniel_hozac: but it should be affected by them, right? (in the case when they are set)
1156371502 M * chris___ ok sorry, i understand now
1156371524 M * daniel_hozac mnemoc: i don't know... it _should_ be.
1156371623 J * Aiken ~james@tooax7-131.dialup.optusnet.com.au
1156371638 M * Bertl welcome Aiken!
1156371648 M * Aiken greetings
1156371667 M * daniel_hozac mnemoc: it seems not, at least not RSS.
1156371735 M * Bertl daniel_hozac: please elaborate?
1156371743 M * daniel_hozac Bertl: tmpfs.
1156371747 M * chris___ daniel_hozac, ok did that, uname -a in the vm reports 2.4.32, but java still crashes while running tomcat during boot
1156371774 M * daniel_hozac Bertl: files created there do not count against the guest's RSS limit, and with secure_remount you can increase the size of the filesystem.
1156371778 M * Bertl tmpfs is separate from memory accounting IIRC
1156371802 M * Bertl you could use dlimits to limit it?
1156371804 M * chris___ i think nptl might be the same problem, i'll try to get java to use green threads
1156371829 M * daniel_hozac Bertl: really? can you set the disk limit on just any tmpfs and it applies to all of them?
1156371836 M * matti Bertl: :))
1156371842 M * Bertl chris___: showing a 2.4 kernel version should disable NPTL
1156371853 M * Bertl daniel_hozac: each tmpfs mount should be handled separately
1156371862 M * chris___ even though redhat puts ntpl support into their 2.4 kernels?
1156371864 M * Bertl daniel_hozac: i.e. have a separate superblock
1156371871 M * daniel_hozac Bertl: ok, that's what i thought.
1156371985 M * daniel_hozac chris___: good point, try 2.4.19.
1156372222 M * chris___ ok, i saw it start multiple processes so it's using green threads, but it still crashes :-(
1156372266 M * chris___ i'll try to start it in rc.local
1156372267 M * harry ftp://ftp.kernel.org/pub/linux/kernel/v2.6/ChangeLog-2.6.17.11
1156372277 M * harry wiiiiiii... www.kernel.org can't even keep up :s
1156372283 M * Bertl daniel_hozac: okay, this is my final analysis, anything I missed? http://paste.linux-vserver.org/283
1156372313 M * Bertl harry: new exploits?
1156372328 M * harry DoS's at least
1156372334 M * harry don't know about sploits
1156372380 M * chris___ must... kill... bill... joy...
1156372717 M * daniel_hozac Bertl: nothing i can think of.
1156372735 M * chris___ ok, i'm making progress. if i start it later with the same env, it dies
1156372768 M * Bertl ok, so it _is_ the env
1156373334 M * chris___ no, unfortunately not, it was the nproc limit
1156373349 M * chris___ i've changed it and i'm back to the old behaviour
1156373382 M * Bertl and it works quite fine when you ssh into the guest and start it from there?
1156373399 M * chris___ well, it works when i start it from vserver enter
1156373407 M * chris___ i haven't tried with ssh yet
1156373411 M * Bertl hmm, let's try with ssh then
1156373493 M * chris___ yep, works with ssh too
1156373666 M * daniel_hozac Bertl: btw, we need to bump VCI_VERSION by one before 2.0.2 as well, right?
1156373686 M * chris___ ok, now it works when i strace catalina.sh in the initscript
1156373691 M * chris___ looks like a heisenbug :-)
1156373826 M * daniel_hozac Bertl: hmm, stable didn't get vc_vci_config?
1156373979 M * daniel_hozac http://vserver.13thfloor.at/Experimental/delta-vcvci-feat01.diff seems harmless to me.
1156374013 M * chris___ ok, i've had enough of this bug, i'm going to bed :-(
1156374020 M * chris___ night all
1156374024 M * daniel_hozac night.
1156374035 M * chris___ thanks for all your help!
1156374056 M * daniel_hozac that's actually a good idea... i'm gonna go to bed as well. good night!
1156374077 Q * chris___ Quit: zZzZ... bugs... zZzZ
1156374280 P * stefani I'm Parting (the water)
1156374443 M * Bertl right, second (or third?) that :)
1156374463 M * Bertl have a good one everyone! cya tomorrow!
1156374471 N * Bertl Bertl_zZ
1156375728 Q * brc_ Ping timeout: 480 seconds
1156375857 J * brc_ bruce@201.19.139.250