1159833607 M * Nei somehow this is realy confusin me 
1159833618 M * Nei the init script talks about /etc/postgresql
1159833691 M * Nei I guess its more likely that debian is screwing me here than vservers
1159833722 M * Bertl hmm, could be ...
1159833803 P * stefani I'm Parting (the water)
1159833804 M * Hurga ubuntu has been screwing me too, today.
1159833838 M * Hurga replaced my vserver kernel (from uniklu) with a generic one... 
1159833864 M * Bertl that was fun for sure :)
1159833918 M * Nei ok it seems that the reason its silently failing are indeed missing config files
1159833927 M * Nei now the question is where are they >_<
1159833929 A * Hurga is still tinkering with apt and dpkg to find out how to detect and prevent something like that the next time.
1159834011 M * doener Hurga: package replacement or just the grub entry?
1159834018 M * doener (or lilo)
1159834020 M * Hurga package replacement
1159834063 M * Hurga both available packages are called "linux-image-2.6.15-27-k7"
1159834083 M * doener ah well, that's "unfortunate"
1159834121 M * Hurga I'd use other words, but - yes.
1159834159 M * _are_ well, why do you name your vserver kernel like some ubuntu kernel?
1159834188 M * Nei mysterious
1159834195 M * Nei I purged the package and installed it again
1159834197 M * _are_ add e.g. the vserver version to the name and it won't break unless you just told it to install any 2.6 kernel
1159834207 M * Nei now I have my config files and it starts as it should -.-
1159834209 M * Hurga _are_: It's a pre-built vserver kernel from http://www2.uni-klu.ac.at/support/VServer
1159834227 M * Nei I hate such problems c.c
1159834229 M * Hurga _are_: So I guess you'd need to ask these guys.
1159834284 M * _are_ he, ok, I build them myself and seems now I even have a valid reason to do so except just 'because'
1159834314 M * Nei put the kernel on hold
1159834332 M * Nei maybe you could also pin it higher but I dont know if that'd be effective
1159834336 M * Hurga _are_: Yes, indeed.
1159834394 M * Nei if you put it on hold the drawback is that it wont get updated automatically, but the benefit is that you have to manually run yourself into troubles ;)
1159834406 M * _are_ Nei: state in your /etc/postgresql/postgres.conf something like
1159834423 M * _are_ syslog = 2                      # range 0-2; 0=stdout; 1=both; 2=syslog
1159834423 M * _are_ syslog_facility = 'LOCAL0'
1159834423 M * _are_ syslog_ident = 'postgres'
1159834436 M * doener Hurga: I'd check apt_preferences(5)
1159834449 M * _are_ (older postgres at my place, but shoudl work nonetheless)
1159834456 M * Nei _are_ thanks for your help, the problem was that debian didnt create my postgres config files for whatever reason
1159834457 M * doener Hurga: probably "origin" can help you, but I didn't look into the details
1159834462 M * _are_ and most likely you find something about 127.0.0.1
1159834482 M * _are_ ah, thought the problem persisted after the config files had been tehre
1159834482 M * Bertl Nei: you do not use xid tagging with dynamic contexts by any chance?
1159834503 M * Nei so a purge & reinstall helped
1159834517 M * Nei Bertl no, that sounds dangerous
1159834536 M * Bertl yep, indeed it is ...
1159834593 Q * samueltc Quit: BitchX-1.1-final -- just do it.
1159834598 M * Nei well you could put a safety block into the xid tagging when the cid is 'too high' 
1159834656 M * _are_ ah, i code my local ip into the cid, so please don't
1159834876 M * doener the "upper" xids are for dynamic use only atm, so nothing would change for you
1159837138 J * node ~node@c-69-143-148-254.hsd1.md.comcast.net
1159837231 J * FCOJ ~mordur@dsl-201-4.hive.is
1159837414 Q * FCOJ 
1159837428 J * FCOJ ~mordur@dsl-201-4.hive.is
1159837499 Q * FCOJ 
1159837518 J * FCOJ ~mordur@dsl-201-4.hive.is
1159837570 Q * FCOJ 
1159837585 J * FCOJ ~mordur@dsl-201-4.hive.is
1159837616 Q * FCOJ 
1159838129 J * id23 ~id@p5081238B.dip0.t-ipconnect.de
1159838139 M * id23 gruetzi #vserver
1159838161 M * id23 2.6.18 is running fine - no problems so far
1159838506 Q * hardwire Quit: Ex-Chat
1159838575 M * Bertl id23: good to hear!
1159839135 Q * Johnnie Remote host closed the connection
1159839589 Q * id23 Remote host closed the connection
1159839749 J * marcfi1 ~mef@c-68-39-177-97.hsd1.nj.comcast.net
1159839779 M * Bertl welcome marcfi1!
1159839803 M * marcfi1 bertl: ugh... looks like I forgot to log off at work.
1159839821 Q * marcfiu Remote host closed the connection
1159839837 M * marcfi1 be right back
1159839844 N * marcfi1 marcfiu
1159839859 M * marcfiu better :)
1159839934 M * marcfiu Has anyone used the cow-links feature in combo with some sort of file-based backup systems, such as rdiff-backup, to further compress away redundancy between separate hosts?
1159840087 Q * DreamerC_ Quit: leaving
1159840105 J * DreamerC ~dreamerc@61-217-226-88.dynamic.hinet.net
1159841615 Q * nokoya Ping timeout: 480 seconds
1159841654 Q * ensc Killed (NickServ (GHOST command used by ensc_))
1159841664 J * ensc ~irc-ensc@p54B4D499.dip.t-dialin.net
1159841736 M * micah does anyone know why my /usr/lib/libvserver.{a,so} files don't contain the vc_new_s_context() call?
1159841744 M * micah is there a config option I'm missing?
1159842006 M * doener micah: did you enable compat and or legacy apis?
1159842246 M * doener good night
1159842364 Q * Blissex Remote host closed the connection
1159842809 M * micah doener: mmm, I didn't have --enable-apis=NOLEGACY configured, if thats what you mean?
1159843087 M * micah apparantly with that enabled they show up
1159843102 M * Bertl sounds normal to me
1159843144 M * Bertl as vc_new_s_context() is legacy (in the kernel terminology :)
1159846785 J * Aiken__ ~james@tooax6-205.dialup.optusnet.com.au
1159847110 Q * Aiken Ping timeout: 480 seconds
1159847933 M * Bertl okidokili, I'm off to bed now ... have a nice one everyone! cya!
1159847941 N * Bertl Bertl_zZ
1159848123 Q * starlein Ping timeout: 480 seconds
1159849160 Q * matled Read error: Connection reset by peer
1159849424 J * matled ~matled@85.131.246.184
1159849740 M * micah anyone know where this went: http://oldwiki.linux-vserver.org/CPU+Scheduling
1159849877 Q * Hurga Remote host closed the connection
1159852551 P * marcfiu 
1159852763 Q * Piet_ Quit: :tiuQ
1159853259 Q * node 
1159856049 J * nokoya ~young@hi-230-82.tm.net.org.my
1159857505 Q * mire Ping timeout: 480 seconds
1159858006 J * ntrs_ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com
1159858021 J * mire ~mire@97-166-222-85.COOL.ADSL.VLine.verat.net
1159858088 Q * ntrs Read error: Connection reset by peer
1159860390 Q * ntrs_ Read error: Connection reset by peer
1159860397 J * ntrs ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com
1159860441 J * prae ~Benjamin@5-63.206-83.static-ip.oleane.fr
1159860600 J * meandtheshell ~markus@85-124-36-168.dynamic.xdsl-line.inode.at
1159860661 Q * nokoya Quit: changing servers
1159861038 J * ntrs_ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com
1159861038 Q * ntrs Read error: Connection reset by peer
1159861215 J * ntrs ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com
1159861215 Q * ntrs_ Read error: Connection reset by peer
1159862625 J * ntrs_ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com
1159862625 Q * ntrs Read error: Connection reset by peer
1159863000 Q * nayco_work Quit: Lost terminal
1159863782 Q * prae Quit: Quitte
1159864030 J * sacha ~sacha@195.225.201.186
1159864039 Q * bon Quit: leaving
1159864150 Q * Aiken__ Ping timeout: 480 seconds
1159864272 J * dna_ ~naucki@p54BCE1A7.dip.t-dialin.net
1159864388 M * Loki|muh derjohn: why does your bind9-capcheck-package not contain a /etc/init.d/bind9 file?
1159864445 J * prae ~Benjamin@5-63.206-83.static-ip.oleane.fr
1159864916 J * nokoya ~young@hi-230-82.tm.net.org.my
1159865706 Q * prae Ping timeout: 480 seconds
1159866354 J * FCOJ ~mordur@dsl-149-109-85.hive.is
1159866383 M * Loki|muh derjohn: but in spite of this, if there is a initscript, your sid package for bind9 works with testing
1159866406 J * yarihm ~yarihm@whitehead2.nine.ch
1159866554 Q * shedi Quit: Leaving
1159866648 J * prae ~Benjamin@5-63.206-83.static-ip.oleane.fr
1159867547 J * shedi ~siggi@dsl-149-109-85.hive.is
1159868603 Q * FCOJ Quit: Leaving
1159868617 M * meandtheshell what's the difference in behaviour of "/etc/vservers/vserver-name/scripts/post-start.d/script" and "/etc/vservers/vserver-name/scripts/post-start" for example? - the same question arises about post-stop, pre-start etc. as well ... TIA for answering!
1159868620 J * FCOJ ~mordur@dsl-149-109-85.hive.is
1159868667 M * meandtheshell I'm asking because as of now I'm reading http://wiki.linux-vserver.org/VServer_Configuration and so ...
1159868694 M * meandtheshell maybe I've overlooked something *shrug*
1159870077 J * coocoon ~coocoon@dslb-084-056-189-206.pools.arcor-ip.net
1159870101 M * coocoon hello
1159870124 M * daniel_hozac meandtheshell: none.
1159870134 M * daniel_hozac post-start.d is a bunch of scripts, post-start is just one script.
1159870143 M * daniel_hozac meandtheshell: oh, and please don't update that page.
1159870158 M * daniel_hozac it's auto-generated from the XML, so i'd really prefer patches to the XML instead.
1159870211 M * daniel_hozac hi coocoon.
1159870836 M * meandtheshell daniel_hozac: ok - I see (the patch thing)
1159870917 Q * doener Ping timeout: 480 seconds
1159870943 M * meandtheshell daniel_hozac: well, I didn't know that in this case it's better to provide a patch to the XML data file - how could I've known ... 
1159870959 M * daniel_hozac meandtheshell: the flower page is generated from the same file ;)
1159870971 Q * matled Ping timeout: 480 seconds
1159870972 M * daniel_hozac (doc/configuration.xml)
1159870974 M * meandtheshell I guess there should be a note somewhere :)
1159870989 M * daniel_hozac yep, i was thinking that too.
1159870994 M * meandtheshell :)
1159871004 M * phedny and maybe the file should be locked for manual editing?
1159871024 M * meandtheshell and furthermore the wiki page should be looked for editing - no?
1159871027 M * daniel_hozac but how would i update it then? :)
1159871045 M * meandtheshell well you should gained rights to do so ...
1159871055 M * meandtheshell hm ...
1159871087 M * meandtheshell by the way - lunch's waiting - back in about 20 minutes or so ... yummy attack :)
1159871112 J * matled ~matled@85.131.246.184
1159871160 J * doener ~doener@host.magicwars.de
1159871368 J * weeble ~weeble@81.52.144.1
1159872645 M * meandtheshell daniel_hozac: hehe funny - you can find that: "This page is automatically created by XSL transformation of configuration.xml. Additions, corrections, etc. should be made in that file, not in the Wiki page." AT THE BOTTOM at the wiki page. Most people (me including) start reading at THE TOP :-)
1159873136 M * fs hello
1159873198 M * waldi meandtheshell: thats why such pages needs access permissions
1159873304 M * fs I have a problem with 2.6.18 and vserver, as soon as an application in a vserver hogs the CPU, the mouse is not usable anymore, it hangs and kumps on movement 
1159873422 M * daniel_hozac fs: and that doesn't happen with a cpuhog running in the host?
1159873447 M * daniel_hozac meandtheshell: oh, hehe, i'll move it to the top.
1159873448 M * fs no
1159873477 M * fs if I run for example dnetc --bench in the host, everything is fine, if I run it in the vserver the mouse practically stops moving
1159873496 M * daniel_hozac and dnetc --bench is what you're running in the guest too?
1159873499 M * fs and keyboard and xserver response is lousy
1159873505 M * fs as an example
1159873509 M * fs yes
1159873513 M * daniel_hozac does your guest have a higher priority?
1159873564 M * fs I have a vserver here compiling a kernel with nice 15 and I have problems tiping
1159873626 M * daniel_hozac which vserver branch?
1159873769 M * fs vs2.0.2.1-t8.patch
1159873779 M * fs it's in the stock debian kernel
1159873780 Q * coocoon Quit: KVIrc 3.2.0 'Realia'
1159873959 M * daniel_hozac can you reproduce it with a vanilla kernel?
1159874235 M * fs there are no patches affecting x86_64 where I am seeing this
1159874275 M * fs so it's basically vanilla+vserver 
1159874318 M * daniel_hozac really? cool.
1159874325 M * fs the only patches applied are some new PCI ids 
1159874333 M * fs and the added areca driver
1159875142 M * derjohn Loki|muh, i dont remember ;) Actually it's really old-versioned and is probably vulnerable to the bind9 DoS tom 6.9.06. It's just there for you to look how I patched the control and debian rules file. everything else is from apt-get source bind9 AFAIR.
1159875346 Q * prae Ping timeout: 480 seconds
1159875911 P * sacha Leaving
1159875974 J * prae ~Benjamin@5-63.206-83.static-ip.oleane.fr
1159877896 J * nayco ~nayco@proxy2.laroche.univ-nantes.fr
1159877915 N * nayco nayco_work
1159877919 M * nayco_work Hello !!
1159878447 J * hkhl ~hkhl@dslb-084-058-199-169.pools.arcor-ip.net
1159878562 P * click [IRSSI]
1159878667 M * shedi in the newest version of vserver, I had defined a loopback interface which disappeared for no reason at all
1159878684 M * shedi then it returned
1159878688 M * shedi :)
1159878973 J * click click@ti511110a080-6047.bb.online.no
1159878975 M * shedi I see there is a newer kernel patch available patch-2.6.17.13-vs2.0.2.1.diff, I'm using patch-2.6.17.11-vs2.02.diff
1159879151 M * daniel_hozac you really should use 2.0.2.1.
1159880139 M * eyck  or 1.2.11rc1 
1159881347 M * daniel_hozac eyck: just to verify i'm not losing my mind, vserver-info - FEATURE migrate; echo $? returns non-zero on your 2.4 hosts, right?
1159881725 M * eyck hmm, vserver-info?
1159881759 M * eyck yes.
1159881760 Q * nayco_work Ping timeout: 480 seconds
1159881825 J * nayco_work ~nayco@proxy2.laroche.univ-nantes.fr
1159882247 M * Borg- aaaaaaaaaaaaaaargh
1159882254 M * Borg- Java is evil.. its evil I'll tell ya
1159882254 M * Borg- ;)
1159882277 M * Borg- anyway.. Is there a hack that I can do ICMP 127.0.0.1 work on guests?
1159882285 M * Borg- unofficial patch? whatever?
1159882315 M * Borg- some braindead folk too is_reachable() method nefore connect()
1159882321 M * Borg- s/nefore/before/
1159882361 M * phedny you make Java connect to 127.0.0.1?
1159882379 M * Borg- phedny: yeah..
1159882391 M * Borg- but before that.. they send ICMP to check if that host is alive..
1159882394 M * phedny and from a non-Java app connecting is possible?
1159882410 M * Borg- phedny: its not problem w/ connect(127.0.0.1) Java or non..
1159882418 M * Borg- its problem w/ sending ICMP packets to 127.0.0.1
1159882426 M * phedny okay
1159882428 M * eyck Borg-: there is a hack for that, but I thought that was incorporated into mainline
1159882431 M * Borg- because normaly.. guest doesnt have lo working
1159882444 M * phedny so you include lo as interface?
1159882450 M * Borg- eyck: hm?? with version?
1159882470 M * Borg- phedny: then bind() 127.0.0.1 will make it visible on all guests + host.. and I dont want to
1159882479 M * Borg- or wait.
1159882486 M * Borg- yeah..
1159882506 M * Borg- I want to have few virtual build servers.. to do different testcases
1159882524 M * Borg- most of stuff work.. except that bloody is_reachable() thingie..
1159882547 M * Borg- I want to check if there is a solution to that. before killing my developers ;) and fixing sources..
1159882659 M * Borg- so.. any ideas? :)
1159882696 M * eyck killing the developers would be more efficient and effective
1159882716 M * Borg- ;D
1159882761 M * eyck hmm, but I'm quite lost as to where the problem is - you want ping working in guest, or maybe you ain't got 'lo' up in guest (which would cause some other additional problems)
1159882944 M * eyck if you want hack for icmp, look for "VXC_RAW_ICMP Context Capability."
1159883056 Q * yarihm Ping timeout: 480 seconds
1159883093 Q * FCOJ Quit: Leaving
1159883214 M * daniel_hozac thanks eyck.
1159883310 M * eyck it says '1'
1159883322 M * daniel_hozac as it should :)
1159883403 M * Borg- ping is working on guest.. but not ping 127.0.0.1 :)
1159883408 M * Borg- any other IP works
1159883547 M * Borg- eyck: okey ;) thx.. Ill try
1159883642 M * Borg- eyck: im just reading sources of vserver-patch
1159883695 M * Borg- +	if ((protocol == IPPROTO_ICMP) && vx_ccaps(VXC_RAW_ICMP))
1159883696 M * Borg- +		goto override; 
1159883705 M * Borg- whats that VCX_RAW_ICMP ? 
1159883712 M * Borg- sth I can enable ?
1159883716 M * Borg- somewhere?
1159883722 M * eyck hack that makes ping work in guests
1159883723 M * Wonka yes
1159883726 Q * prae Ping timeout: 480 seconds
1159883735 M * eyck without enabling _NET_RAW CAP
1159883748 M * eyck if ping is working, then you probably just need lo
1159883760 M * Borg- hmm this is weird then
1159884029 M * Wonka why isn't the damn great flower page linked in the new wiki?
1159884035 M * Wonka or, if it is, where is it?
1159884065 M * daniel_hozac because if you google on great flower page, it's the first hit?
1159884078 M * Borg- eyck: no no.. adding lo iface is wrong
1159884083 M * daniel_hozac and there's also an alternative wiki-ized version at http://linux-vserver.org/VServer_Configuration
1159884085 M * Borg- (w/ ip 127.0.0.1)
1159884104 M * Borg- anyway.. I'll just see /net/ipv4/af_inet.c 
1159884106 M * daniel_hozac and i'm positive it's linked, somewhere.
1159884110 M * Wonka Borg-: look for /etc/vservers/$name/flags at the big flower page
1159884119 M * Borg- later.. maybe there is sth I can add ;)
1159884145 M * daniel_hozac flags doesn't have anything to do with ccaps though...
1159884148 M * mnemoc it's not the first if you look for vserver weed page :\
1159884178 M * daniel_hozac mnemoc: why would you expect it to be? that's not what it's called :)
1159884185 M * mnemoc :D
1159884202 M * eyck not having lo is wrong.
1159884207 M * daniel_hozac eyck: having lo is wrong.
1159884213 M * Borg- Wonka: there is nothing about ICMP or LO
1159884214 M * daniel_hozac eyck: at least until it's virtualized.
1159884217 M * eyck daniel_hozac: why?
1159884218 M * eyck oh, 
1159884223 M * daniel_hozac Borg-: see the ccapabilities link.
1159884226 M * eyck still, not having lo is wrong, 
1159884230 M * daniel_hozac eyck: why?
1159884236 M * daniel_hozac lo will be used for all local packets anyway.
1159884251 M * eyck having lo not virtualized might be worse in some situations
1159884278 M * Borg- daniel_hozac: okey. I see sth :) lets test it
1159884284 M * daniel_hozac virtualized lo should be available RSN, along with IPv6 support ;)
1159884292 M * daniel_hozac Borg-: shouldn't change anything though, as it's given by default.
1159884305 M * Borg- daniel_hozac: whats what im aware of... :/
1159884314 M * Borg- daniel_hozac: I see sth interesting in /net/ipv4/af_inet.c
1159884325 M * Borg- but currently im too busy to check it out :/// maybe later..
1159884332 M * Borg- damn boss gives me some shitty work to do.. argh
1159884399 J * prae ~Benjamin@5-63.206-83.static-ip.oleane.fr
1159884823 J * Hurga nobody@p508A88A7.dip0.t-ipconnect.de
1159885344 J * FireEgl FireEgl@Sebastian.Atlantica.US
1159885545 M * Wonka daniel_hozac: what's missing for lo and v6?
1159885733 M * eyck hmm, I don't get one thing... with NGNET we were supposed to get lo virtualisation... what else did we need NGNET for?
1159885956 M * Borg- well.. lo virtualization is not so necessary
1159885980 M * Borg- as far as we can fully rewrite packets (IP stack) from 127.0.0.1 to first guest address
1159886325 M * eyck hmm, I don't think that's very nice
1159886347 M * doener it breaks the "only this box/vserver can see it" assumption
1159886348 M * eyck things that want to talk via 127.0.0.1 usually are security sensitive, 
1159886369 M * eyck and you don't want such data willi-nilly rewritten to some public ip
1159886626 Q * prae Ping timeout: 480 seconds
1159887255 N * Bertl_zZ Bertl
1159887259 M * Bertl morning folks!
1159887268 M * h01ger morning Bertl :)
1159887299 M * daniel_hozac Wonka: being merged into devel ;)
1159887320 M * Wonka daniel_hozac: mh.
1159887331 M * daniel_hozac morning Bertl
1159887375 M * romke morning
1159887445 M * meandtheshell Bertl: morning
1159887748 J * prae ~Benjamin@5-63.206-83.static-ip.oleane.fr
1159888231 Q * prae Ping timeout: 480 seconds
1159888467 J * node ~dwindsor@linux2.gl.umbc.edu
1159888788 M * matled how can I make nc -l -s 127.0.0.1 1234 and then nc -v 127.0.0.1 1234 get to work inside a jail?
1159888960 M * Borg- matled: you mean inside a guest?
1159888965 M * Borg- it should work out of the box
1159888965 M * matled yep
1159888969 M * matled mhh
1159888976 M * Borg- 127.0.0.1 is rewrited to first guest IP
1159888977 M * matled I think I stopped the iptables script
1159888994 J * prae ~Benjamin@5-63.206-83.static-ip.oleane.fr
1159888996 M * matled yep, all tables are empty
1159889012 M * matled # nc -v 127.0.0.1 1234
1159889012 M * matled localhost [127.0.0.1] 1234 (?) : Connection refused
1159889035 M * matled under problematic programs I found someone suggesting to create a lo interface inside the vserver
1159889120 M * Borg- dont do that
1159889131 M * Borg- wait a moment please.. Ill check that on my guest
1159889158 M * Borg- oh sorry.. I cant :) but it works 100%
1159889177 M * Borg- ptw
1159889184 M * Borg- propeely netcast listener is
1159889190 M * matled argh!
1159889192 M * Borg- nc -l -p 1234 127.0.0.1
1159889194 M * Borg- and connect
1159889200 M * Borg- nc 127.0.0.1 1234
1159889204 M * matled I missed the -p :/
1159889207 M * Borg- :PP
1159889220 M * matled bind(3, {sa_family=AF_INET, sin_port=htons(0), sin_addr=inet_addr("127.0.0.1")}, 16) looks a bit strange
1159889287 M * Borg- but works.. anyway.. going home.. cu
1159889383 J * stefani ~stefani@tsipoor.banerian.org
1159889553 M * Bertl matled: is that the control connection, yes?
1159889561 Q * prae Ping timeout: 480 seconds
1159889585 M * Bertl matled: basically you can do two tings:
1159889607 M * Bertl - assign a different lo address to the guest, e.g. 127.0.0.42
1159889619 M * Bertl - change the bind config to use the first ip
1159889726 M * matled Bertl: I just missed the -p parameter to nc -l and it defaults to 0
1159889890 M * Bertl ah, okay
1159889926 M * matled I'm trying to run ejabberd inside a vserver and tried to do something like it does when it fails
1159889994 M * eyck ejabberd works fine here...
1159890026 M * matled I just created a debian etch vserver and installed the package
1159890101 M * matled mh, k, now it works
1159890238 M * matled ah, iptables.. I thought I tried it without the firewall enabled
1159890800 J * prae ~Benjamin@5-63.206-83.static-ip.oleane.fr
1159891346 Q * prae Ping timeout: 480 seconds
1159892005 J * bonbons ~bonbons@83.222.36.111
1159892089 J * prae ~Benjamin@5-63.206-83.static-ip.oleane.fr
1159892800 Q * ||Cobra|| Read error: No route to host
1159892980 J * ||Cobra|| ~cob@pc-csa01.science.uva.nl
1159893729 J * jujux ~jujux@lns-bzn-42-82-255-77-41.adsl.proxad.net
1159893817 Q * jujux 
1159893940 J * jujux ~jujux@lns-bzn-42-82-255-77-41.adsl.proxad.net
1159893949 M * jujux hello
1159893967 M * daniel_hozac hi
1159894037 M * jujux I'm testing vserver with etch, and ask me about vserver in vserver. How secure is this solution ?
1159894089 M * jujux And how valid/viable is it ?
1159894108 M * daniel_hozac guests within guests is not possible.
1159894132 M * jujux ok. Why ?
1159894159 M * jujux if the answer is to technical, nevermind
1159894200 M * daniel_hozac probably because noone wanted it bad enough yet :)
1159894220 M * phedny jujux: you could try vserver-in-vmware, that should work ;)
1159894234 M * daniel_hozac vserver-in-xen should work as well.
1159894255 M * phedny daniel_hozac: Xen doesn't work with modified kernel?
1159894259 M * jujux xen and vmware aren't solutions for me, thanks for suggestions
1159894280 M * daniel_hozac phedny: Debian supplies prebuilt xen+vserver kernels.
1159894286 M * daniel_hozac as do i, for Fedora.
1159894330 M * phedny well, I only read about Xen, never really tried it, so you should know better :)
1159894416 M * jujux daniel_hozac: I'm sorry, but i'm not sure to understand the sense of "none wanted it bad enough"
1159894451 M * daniel_hozac jujux: what's the point of implementing features nobody will use?
1159894466 M * jujux ok
1159894494 M * daniel_hozac what would you be using guests within guests for?
1159894508 J * hardwire ~sspencer@209.112.165.133
1159894543 M * jujux too have a kind a master admin for all vservers, in a vserver, for example security audit and archives
1159894565 M * daniel_hozac and why not use the host for that?
1159894567 M * jujux kind of
1159894680 M * jujux is that clear ? :-/
1159894696 M * daniel_hozac what?
1159894700 M * jujux I'm sorry, i will be back in 5 min
1159895041 M * jujux I think I'm deconnected, if not I just reconnect
1159895055 Q * jujux Quit: Leaving
1159895072 J * jujux ~jujux@lns-bzn-42-82-255-77-41.adsl.proxad.net
1159895136 M * jujux daniel_hozac: did you answer me after that hardwire joined the channel ?
1159895138 M * hardwire I wish I were a studmuffin
1159895153 M * daniel_hozac 18:56 < daniel_hozac> and why not use the host for that?
1159895292 M * Bertl hardwire: why is that so?
1159895325 M * hardwire Bertl: dpm
1159895326 M * hardwire err
1159895327 M * hardwire heh
1159895338 M * hardwire nothing.. in the mornings I just say the first random thing out of my brain
1159895369 M * Bertl i.c. well :)
1159895450 M * Borg- holy shit..
1159895455 M * Borg- jujux is babling..
1159895486 M * cehteh jujux: you can nest the filesystems of guest, that makes nesting somewhat possible
1159895489 M * Bertl okay, have to run now .. back in the evening ...
1159895493 N * Bertl Bertl_oO
1159895927 Q * prae Quit: Quitte
1159895929 P * stefani I'm Parting (the water)
1159896783 P * jujux Leaving
1159896817 J * jujux ~jujux@lns-bzn-42-82-255-77-41.adsl.proxad.net
1159896817 Q * jujux Quit: Leaving
1159896828 J * jujux ~jujux@lns-bzn-42-82-255-77-41.adsl.proxad.net
1159896832 M * jujux sorry...
1159896896 M * jujux daniel_hozac (?): I would prefer to not install security tools on the host
1159896896 M * jujux I saw that others persons said my nick
1159896960 M * jujux and I don't understand babling
1159897039 M * jujux someone said something about the fs (?) , or maybe something else
1159897363 M * meandtheshell jujux: Linux-VServer does virtualization and isolation - I guess what you want is being able to administer a bunch of isolated servers aka guests. As daniel already mentioned ... you do have one host and n guests. So the administrator that has unlimted access to the host is what you mentioned as "master admin". Inside each guest there might be a administrator as well but that one only controlls a specific guest - no other gu
1159897466 M * meandtheshell to avoid confusion - virtual private server (vps), guest, vserver - different names for the same thing
1159897542 M * jujux ok, thanks for precisions
1159897592 M * meandtheshell no problem - you are welcome - that one http://www.linux-vserver.org/Paper might enlighten you as well 
1159897606 M * jujux I wanted to ask as I readed some todo about vserver in vserver security in a very old vserver weekly news
1159897618 M * meandtheshell jujux: I see
1159897627 M * jujux so thanks to all, for your time, and thanks to developers for this great software
1159897667 M * jujux bye
1159897667 Q * jujux Quit: Leaving
1159897910 Q * hkhl Quit: Chatzilla 0.9.75 [Firefox 1.5.0.7/2006090918]
1159898112 M * Hurga root@kabelmarder:/vservers# rmdir otrs
1159898112 M * Hurga rmdir: otrs: Device or resource busy
1159898149 M * Hurga ex vserver mountpoint (LVM). any ideas? lsof and mount say it's clean
1159898331 M * daniel_hozac do you still have guests running?
1159898339 M * daniel_hozac was the mountpoint there when they were started?
1159898439 M * Hurga I have one other guest, shoud be unrelated
1159898484 M * Hurga I think I added that mount point later. But I've been experimenting a bit with other guests
1159898485 J * pookey ~pookey@is.paranoid.org.uk
1159898504 M * pookey h01ger: thanks ;)
1159898522 M * pookey hi all - vserver currently doesn't support live migration between servers does it ?
1159898534 M * daniel_hozac Hurga: what does vnamespace -e <second guest> cat /proc/mounts say?
1159898538 A * h01ger waves at pookey :)
1159898540 M * daniel_hozac pookey: no.
1159898550 M * pookey daniel_hozac: is it planned ?
1159898577 M * daniel_hozac not that i'm aware of.
1159898636 M * pookey other hten openVZ, are there any other implementations like this that do suppport migration ?
1159898660 M * Hurga daniel_hozac: <second guest> is the name of the only running vserver, yes?
1159898772 M * doener yes
1159898822 M * Hurga long list, and I have referneces to m busy mount point, indeed...
1159898853 M * Hurga what's the difference to just executing  cat /proc/mounts inside the vserver?
1159898935 M * daniel_hozac /proc/mounts inside the guest hides the ones outside of the chroot.
1159898957 M * Hurga ah, I see.
1159898984 M * daniel_hozac vnamespace -e <second guest> umount ... should free it.
1159899185 M * Hurga Thanks! 
1159899215 M * Hurga Is that in the docs, and where would I have had to look for it?
1159899688 M * daniel_hozac http://oldwiki.linux-vserver.org/Namespaces
1159900224 J * sp sp@delta.sp.or.at
1159900732 Q * hardwire Quit: Ex-Chat
1159900763 J * hardwire-afk ~hardwire@89-208-58-66.gci.net
1159900766 N * hardwire-afk hardwire
1159900781 M * daniel_hozac mugwump: i fixed my last significant problems with the fai build method, could you retest to see if i broke anything?
1159900788 N * hardwire hardwire-afk
1159900790 N * hardwire-afk hardwire
1159900803 N * hardwire hardwire-afk
1159900806 N * hardwire-afk hardwire
1159901006 P * pookey 
1159901655 Q * shedi Quit: Leaving
1159902011 M * daniel_hozac ensc: ping
1159902836 M * mugwump daniel_hozac: will test today
1159903331 M * daniel_hozac note that it will require the secure-mount too, so just updating the script won't do it ;)
1159903405 M * mugwump sure.  sadly also debootstrap won't work without mknod
1159903434 M * daniel_hozac oh, so we need to add that cap to the vattribute call?
1159903436 J * cdrx ~legoater@cap31-3-82-227-199-249.fbx.proxad.net
1159903464 M * mugwump it's only required for running fai-setup
1159903487 M * mugwump the actual install doesn't use debootstrap, it just untars a .tgz
1159903492 M * daniel_hozac ok.
1159903524 M * mugwump actually, I guess it still needs to be able to make /dev/null etc
1159903538 M * daniel_hozac yes, i'd say so.
1159903547 M * daniel_hozac although we could let util-vserver handle that.
1159903555 M * daniel_hozac (like the template build method does)
1159903557 M * mugwump I remember doing some partial tar extract code in the fai modifications, I'll look into it again once I get to work
1159903598 M * mugwump yes, if it can make the essential /dev nodes up front, that would make it easy
1159903660 M * daniel_hozac populateDev creates the default set.
1159903670 M * mugwump ok
1159904088 J * Blissex ~Blissex@82-69-39-138.dsl.in-addr.zen.co.uk
1159904871 Q * kevinp Quit: Compiling...
1159904903 J * Term_ ~user1@189.131.8.134
1159905310 Q * _are_ Ping timeout: 480 seconds
1159905313 J * _are_ are@dslb-084-057-143-111.pools.arcor-ip.net
1159905603 M * micah h01ger: as far as I know vserver-debianutils is broken (according to everyone here), and all i know that they do is they make a vserver by exluding some obvious packages
1159905613 M * micah I haven't tried it recently to really evaluate
1159905621 J * shedi ~siggi@inferno.lhi.is
1159905635 M * micah and Ola hasn't bothered to put it in the pkg-vserver svn repo
1159905653 M * micah i conflicted with it in the backports version, and someone complained
1159905713 Q * Blissex Remote host closed the connection
1159905714 M * micah on another note, I am going to ask for kernel-patch-vserver to be removed from the archive, as its just another package that would need to be maintained for security, and doesn't have that big of a value now
1159906079 M * micah h01ger: ^^
1159906257 M * matled should the saving from hashify be visible in df?
1159906337 M * daniel_hozac not immediately.
1159906338 M * matled ah, it does now
1159906349 M * daniel_hozac you have to restart the guests before it's completely visible.
1159906756 Q * bonbons Quit: Leaving
1159906783 M * h01ger micah, if its broken we should file a serious bug, so they are not shipped in etch
1159906815 M * micah I've got some limits set for a vserver, but it seems like it is still able to take over the box -- I had to reboot the host box because I couldn't fork another process to ssh in to fix things
1159906837 M * micah h01ger: I agree, however we need someone to figure out what specifically is broken, Bertl_oO might have an opinion
1159906838 M * daniel_hozac micah: what limits do you have set? are they low enough to save the host?
1159906900 M * micah oh wow
1159906908 M * micah it seems like I do *not* have limits set
1159906940 M * micah I must have neglected to copy those from the previous host it was on, I normally would have just rsync'd the /etc/vservers/<vserver> directory, but apparantly... I didn't
1159907040 M * micah daniel_hozac: I *will* have 'sched_prio' set, and /etc/vservers/paloma/schedule set with '20 40 500 200 4800 dummy' (all on different lines of course)
1159907062 M * daniel_hozac sched_prio is still just priorities...
1159907099 M * daniel_hozac and not being able to fork a new process sounds more like RAM.
1159907110 M * daniel_hozac (alternatively process limits)
1159907141 M * micah good point
1159907163 M * micah rlimits would be how I would set memory stuff if I am correct
1159907171 M * daniel_hozac yep.
1159907178 M * daniel_hozac and process limits.
1159907190 A * micah sets 'as', 'memlock', 'rss'
1159907298 J * ptl_xid ~iuuuju@187-200-231-201.fibertel.com.ar
1159907310 M * ptl_xid Hi to all !
1159907328 M * daniel_hozac hello
1159907338 M * ptl_xid Hi Daniel
1159907374 M * ptl_xid I'm having an strange issue building a vserver... would you like to hear it ?
1159907388 M * daniel_hozac yes.
1159907429 M * ptl_xid ok
1159907450 M * ptl_xid I'm using the same command that I always used before: 
1159907451 M * ptl_xid vserver vs106 build -n vs106 --hostname vs106 --interface eth0:192.168.0.4/24 -m debootstrap -- -d etch
1159907474 M * daniel_hozac -n vs106 isn't needed.
1159907497 M * ptl_xid yeah, you´re right !
1159907510 N * Bertl_oO Bertl
1159907514 M * Bertl evening folks!
1159907519 M * ptl_xid Hi Bertl
1159907529 M * Bertl daniel_hozac: this -n comes from the strange help output
1159907563 M * daniel_hozac i suppose...
1159907623 M * ptl_xid ok, the real problem is that the bootstrap process begins, but fails with the following output
1159907642 M * ptl_xid I: Installing base packages...
1159907642 M * ptl_xid W: Failure while configuring base packages.  This will be attempted 5 times.
1159907662 M * ptl_xid and it attempts 4 times more...
1159907693 M * micah ptl_xid: do you have the most current debbootstrap package installed?
1159907708 J * dna___ ~naucki@p54BCE1A7.dip.t-dialin.net
1159907709 M * ptl_xid I'm using etch
1159907738 M * ptl_xid I didn't upgraded it since... 1 month I think... should I ?
1159907746 M * micah uff, these limits aren't helping much, server is slooow
1159907749 M * micah ptl_xid: yes, you should
1159907766 M * Bertl micah: hmm?
1159907768 M * ptl_xid ohh... but it used to work with the version I have
1159907808 M * micah Bertl: oh i just set some limits on a vserver that seems to be taking over this host, restarted and right away its having a hard time just even doing a vtoop
1159907824 M * Bertl what limits?
1159907827 M * micah I think I may not have set strong enough limits, or something else is happening
1159907839 M * micah I set sched_io and rlimits
1159907856 M * Bertl and it is not hogging the cpu by any chance?
1159907857 M * micah it doesn't appear as if swap is being dipped into at all, but load is up to 21
1159907861 M * micah I think it might be
1159907875 M * ptl_xid sorry micah, have a look at this:
1159907876 M * ptl_xid rs01:/vservers/vs106# apt-get install debootstrap
1159907876 M * ptl_xid Reading package lists... Done
1159907876 M * ptl_xid Building dependency tree... Done
1159907876 M * ptl_xid debootstrap is already the newest version.
1159907886 M * micah ptl_xid: did you do apt-get update first?
1159907895 M * ptl_xid yep
1159907926 M * micah ptl_xid: it sounds like your debootstrap is trying to contact a mirror that is either not responding or is broken some other way
1159907943 M * ptl_xid but actually it downloads all packages
1159907946 M * micah Bertl: I think this guest is hogging the cpu
1159907971 M * hardwire speaking of hogging the cpu
1159907974 M * micah ptl_xid: ok, then it probably is a particular package that is not installing properly, and the error should be further up
1159907975 M * hardwire byo I hate zope some days
1159907989 A * micah stops the vserver before it stops the host :)
1159908002 M * ptl_xid micah: the process is failing at the end
1159908006 M * Bertl micah: in this case you want to adjust the scheduler
1159908027 M * daniel_hozac ptl_xid: could you post the entire log?
1159908038 M * micah ptl_xid: don't post the entire log in here
1159908052 M * micah ptl_xid: use a pastebin, ask if you need to know what that is :)
1159908069 M * daniel_hozac just typing paste.linux-vserver.org would've saved you some keystrokes :)
1159908070 M * ptl_xid ok, what do you want... all the output since the vserver build ?
1159908073 M * daniel_hozac yes.
1159908085 M * ptl_xid ok, I will
1159908091 Q * dna_ Ping timeout: 480 seconds
1159908095 M * micah daniel_hozac: heheh
1159908235 M * micah i'm not very good at adjusting the scheduler params
1159908256 Q * doener Ping timeout: 480 seconds
1159908318 N * ptl_xid patulo
1159908374 J * doener ~doener@host.magicwars.de
1159908401 J * dna_ ~naucki@p54BCE1A7.dip.t-dialin.net
1159908583 M * micah hrm, ok that didn't work
1159908667 M * patulo hey micah!, look, I decided to run it again before posting it... and it worked... but I just pressed the ¨up¨ key and used the history... 
1159908705 M * micah patulo: well good! :)
1159908719 M * daniel_hozac patulo: must've been a bad mirror then ;)
1159908786 Q * dna___ Ping timeout: 480 seconds
1159908815 Q * Term_ Quit: ircII2.8.2-EPIC3.004 --- Bloatware at its finest.
1159908850 M * micah i set the cfq scheduler, and either what was causing the load has gone away, or that solved it
1159909262 M * patulo yeah, but look, now I entered the vserver, installed ssh, logged in and I'm getting segfaults... What's happening here ??...
1159909275 M * patulo vs106:~# apt-get upgrade
1159909275 M * patulo Segmentation faultsts... 0%
1159909286 J * dna___ ~naucki@p54BCE1A7.dip.t-dialin.net
1159909418 M * _are_ patulo: mmap stuff?
1159909429 M * micah that sounds like bad memory
1159909450 M * patulo How do I probe it ?
1159909468 M * micah patulo: run memtest86+ on your machine overnight
1159909471 Q * dna_ Ping timeout: 480 seconds
1159909484 M * micah patulo: preferably the one that you boot into via a CD or grub
1159909510 M * _are_ well, easiest solution: you ran in the typical apt cache shortage stuff, you can fix it by adding to your apt.conf: APT::Cache-Limit 33554432;
1159909519 M * _are_ if this fails, *then* check for bad memory
1159909531 M * patulo I'll try it
1159909554 M * micah well, that would be if patulo got the mmap error, but it looks like the segfault was happening before apt even finished generating its package list
1159909573 M * micah patulo: although if there was an mmap error, then that would be it :)
1159909602 M * _are_ micah: basically, yes, but I have seen some occasions when the mmap stuff has not been displayed
1159909662 M * micah ooh thats interesting
1159909679 M * patulo micah: the segfault countinues using the Cache-Limit...
1159909689 M * _are_ then check memory, indeed
1159909716 M * patulo ok, _are_, thank you very much !
1159909750 M * h01ger micah, the package description of vserver-debiantools says: "There are tools to create and duplicate a virtual server. There is also a tool to strip a copy of a normal debian installation from hardware dependent things." - so i guess, not all of this functionality is also in util-vserver, right?
1159909768 M * daniel_hozac h01ger: if not, we should add it.
1159909796 M * meandtheshell hehe - funny - I actually stumbled over the same issue today 
1159909854 M * meandtheshell question: who wrote the tools contained within vserver-debiantools? resp. who takes care of them as of now?
1159909864 M * micah if anyone can offer any scheduler hints, it seems as if the cpu is being hogged by this one guest still: http://paste.linux-vserver.org/428
1159909933 M * micah h01ger: i think it is available in util-vserver, just by setting the --exclude=$REMOVE_PACKAGES on vserver build time?
1159909946 M * micah meandtheshell: Ola does
1159909948 J * dna_ ~naucki@p54BCE1A7.dip.t-dialin.net
1159909966 M * meandtheshell micah: right - I checked as well 
1159909975 M * micah its a confusing package for people
1159909981 M * meandtheshell I agree
1159909998 M * daniel_hozac micah: sched_hard would probably work better, if your kernel supports it.
1159910000 M * micah and reported to be broken here, but I have not been able to convince him to make it go away because he uses it (and apparantly others do as well), but for waht, I am not sure
1159910035 M * micah daniel_hozac: wouldn't that cause problems with processes? like segfaults or somesuch?
1159910043 M * daniel_hozac hmm, why would it?
1159910087 Q * dna_ 
1159910137 M * h01ger micah, duplicate and strip existing installation from hw-dependend things?
1159910157 M * daniel_hozac strip a copy, not strip and copy.
1159910158 M * meandtheshell micah: apt-file list vserver-debiantools | grep sbin
1159910286 M * micah daniel_hozac: I'm not sure, I thought I read that somewhere
1159910303 M * daniel_hozac micah: it really, really shouldn't.
1159910313 M * micah h01ger: well I think the 'strip' just removes packages that are hardware related
1159910319 M * micah h01ger: I dont know about the copy part
1159910346 Q * dna___ Ping timeout: 480 seconds
1159910367 M * h01ger so i tend to think its sensible to have. maybe ask ola to remove the create part and include a pointer, a depends even, to util-vserver?
1159910390 M * daniel_hozac it already depends on util-vserver.
1159910432 M * micah there are four programs it provides: dupvserver, newnfsvserver, newvserver and stripserver
1159910436 M * micah there are manpages as well
1159910481 M * micah the man page for stripvserver is really crazy
1159910486 M * micah its formatted all wrong
1159910490 M * meandtheshell micah:  cat /usr/share/doc/vserver-debiantools/README
1159910492 M * micah and it includes a lot of legacy stuff
1159910496 M * micah meandtheshell: yes,  I know
1159910521 M * meandtheshell :) just because the history stuff there ...
1159910682 M * ntrs_ I have this misterious problem with 2.1.1-rc35 on 2.6.17.13. mysqld is hanging after about a day of uptime, after a restart of the guest it works again for about a day. Has anyone seen anything similar to this?
1159910694 M * ntrs_ daniel_hozac, is there an rpm for 211-rc2?
1159910718 J * Blissex ~Blissex@82-69-39-138.dsl.in-addr.zen.co.uk
1159910721 M * micah daniel_hozac: unfortunately, it appears waldi has # CONFIG_VSERVER_HARDCPU is not set
1159910730 M * daniel_hozac ntrs_: no, not yet...
1159910745 M * ntrs_ ok, does the tar include a spec file so I can build one?
1159910746 M * daniel_hozac micah: still? i thought someone petitioned him to change it :)
1159910759 M * daniel_hozac ntrs_: it does, but you can't build one from it.
1159910806 M * ntrs_ daniel_hozac, why not? so, the only way is configure, make, make install?
1159910825 M * ntrs_ daniel_hozac, should I change any of the configure parameters? paths?
1159910826 M * daniel_hozac the specfile sets the version to 0.30.211-rc2, which is invalid.
1159910846 M * ntrs_ hmm, what is the valid version then?
1159910846 M * daniel_hozac ntrs_: personally i build using the specfile.
1159910855 M * ntrs_ maybe I can change the spec file?
1159910865 M * daniel_hozac i set version to 0.30.211, release to 0.1.rc2.
1159910870 M * ntrs_ ok
1159910874 M * ntrs_ in the spec file?
1159910874 M * meandtheshell daniel_hozac: so you're thinking about integrating those bourne shell scripts from vserver-debiantools into your util-vserver branch - no?
1159910885 M * daniel_hozac ntrs_: right.
1159910888 M * ntrs_ ok
1159910891 M * waldi micah: 2.6.18 is unusable anyway
1159910904 M * ntrs_ anyone seen these hanging issues with 2.1.1-rc35?
1159910916 M * micah waldi: i'm using 2.6.17
1159910928 M * daniel_hozac ntrs_: then just rpmbuild --define '_sourcedir '`pwd` util-vserver.spec
1159910949 M * daniel_hozac ntrs_: there were reports of mysqld hanging.
1159910965 M * ntrs_ daniel_hozac, are you talking about my reports?
1159910977 M * daniel_hozac ntrs_: yes, but someone else was able to reproduce them.
1159910987 M * ntrs_ daniel_hozac, wow, who was that?
1159910998 M * daniel_hozac hmm, derjohn i think, i'd have to check my logs.
1159911000 M * ntrs_ Bertl, are you reading this?
1159911015 M * Bertl sure, I'm all ears (eyes?:)
1159911067 M * ntrs_ Bertl, hehe
1159911109 M * ntrs_ daniel_hozac, let me know when you know for sure if it was derjohn, so I can discuss this with him further.
1159911113 M * micah waldi: so if I can petition to enable CONFIG_VSERVER_HARDCPU in 2.6.17 :)
1159911155 M * daniel_hozac ntrs_: yep, looks like it was derjohn.
1159911174 M * waldi no, 2.6.17 is froen
1159911198 M * ntrs_ daniel_hozac, thanks
1159911208 M * ntrs_ Bertl, are you aware of derjohn's findings?
1159911212 M * ntrs_ derjohn, ping?
1159911236 M * Bertl not really, IIRC he experienced some issues a while ago, but I didn't relate it to your issues
1159911495 M * fs what's the correct method of adding a loopback iface to a vserver? is it supported at all?
1159911522 M * Bertl fs: you do not add or remove interfaces to guests at all
1159911539 M * Bertl fs: networking is based on ip isolation
1159911549 M * daniel_hozac meandtheshell: i wouldn't integrate the scripts themselves, i'd add their functionality.
1159911562 M * fs so it's not possible to bind on 127.0.0.1 inside of a vserver?
1159911567 M * Bertl i.e. if you give a certain ip to a guest, the interface which carries it will become visible
1159911576 M * daniel_hozac fs: it is.
1159911585 M * Bertl fs: bindings to 127.0.0.1 are remapped to the first assigned ip
1159911622 M * fs Bertl: ok, the whole problem: I run a build daemon in a vserver, and some packages do ping 127.0.0.1 or bind to 127.0.0.1 to accomplish some tests on themself, and they currently fail
1159911653 M * Bertl fs: have you adjusted 'localhost' in /etc/hosts yet?
1159911683 M * fs it points to the IP, but not the loopback address
1159911709 M * Bertl okay, that should be fine, still some packages use 127.0.0.1?
1159911714 M * meandtheshell daniel_hozac: sure - that means rewrite it in C - a bunch of work ...
1159911718 M * fs apparently yes
1159911723 M * Bertl fs: note: those packages are broken
1159911728 M * daniel_hozac meandtheshell: no, the utils are mostly bash...
1159911741 M * Bertl fs: but nevertheless, you can assign 127.0.0.1 to the guest too
1159911744 M * fs nagios-plugins for example do fail to build 
1159911749 M * daniel_hozac meandtheshell: a proper clone build method would require a C helper though.
1159911771 M * Bertl fs: if they use hardcoded 127.0.0.1, I would go and change that to 'localhost' wherever possible
1159911776 M * fs perl used to fail unless localhost points somewhere
1159911795 M * Bertl yes, that is fine, i.e. localhost should be the first assigned ip
1159911800 M * meandtheshell I see - I just checked two utils right away lsxid and vkill
1159911885 M * fs if I assign 127.0.0.1 to the guest, it will get shut down in the host too on vserver stop? 
1159911887 M * daniel_hozac meandtheshell: those are the lowest level utils.
1159911904 M * daniel_hozac meandtheshell: vserver is a script that calls other scripts that eventually calls those.
1159911911 M * meandtheshell ok - I see - that is why ...
1159911918 M * Bertl fs: not necessarily, depends on your config
1159911928 M * Bertl fs: usually you would add that with 'nodev'
1159911940 M * fs I see
1159911947 M * fs ok I'll give it a try
1159911971 M * Bertl just make sure that it is the first one :)
1159911998 M * Bertl otherwise you will end up with remapping 127.0.0.1 to something else
1159912019 M * fs and apps binding to 127.0.0.1 in the vserver won't be accessible from outside? =)
1159912029 M * Bertl they will be
1159912036 M * fs uuh
1159912042 M * Bertl after all, you are sharing the host 127.0.0.1 now with the guest
1159912052 M * fs hm that's not the plan
1159912065 M * fs so fixing tests to use localhost is the right way
1159912066 M * Bertl it would be better to fix the broken packages :)
1159912091 M * Bertl after all, you could be perfectly fine with using 127.0.0.42 too :)
1159912105 M * Bertl nobody says you have to use 127.0.0.1 :)
1159912139 M * harry 127.25.216.37 ?
1159912152 M * Bertl yeah, nice one too :)
1159912162 M * harry or: how to ask a silly question about /8 networks :)
1159912181 M * fs I see
1159912185 M * fs thats a good point
1159912190 M * harry .
1159912195 M * harry ^ is a good point :)
1159912211 M * fs will it still be shared among the vservers and the host?
1159912236 M * Bertl networking happens on the host
1159912254 M * Bertl but unless you bind anything on the host, it will not interfere
1159912273 M * fs no I fixed everything to bind specific addresses
1159912302 M * Bertl so that should be fine then
1159912614 M * fs another issue I already reported earlier today: on 2.6.18+vs2.0.2.1-t8, a program inside of the vserver taking CPU memory makes the system unresponsive for interactive tasks like X, the mouse begins to jump and stuck
1159912624 M * fs and typing response is sluggish
1159912650 M * fs s/memory//
1159912651 M * Bertl is hard cpu scheduling enabled?
1159912654 M * fs no
1159912663 M * Bertl then this is kind of expected
1159912672 M * fs on x86_64 both UP and smp
1159912707 M * Bertl probably the same happens when you start that CPU hog on the host
1159912711 M * fs no
1159912714 M * Bertl no?
1159912721 M * fs only if started in the vserver
1159912728 M * Bertl that's indeed interesting
1159912738 M * Bertl could you check the priorities for me there?
1159912751 M * fs yes
1159912773 M * Bertl maybe start two test cpu hogs (one on the host and one inside the guest)
1159912791 M * Bertl they should get the same amount of cpu and identical priorities
1159912799 M * fs let me boot my testbox into the kernel
1159912812 M * Bertl (assumed that prio bias is set to 0 and sched_prio is disabled)
1159912942 M * micah maybe cfq needs to be set?
1159912958 M * Bertl cfq should be unrelated to cpu bound hogs
1159912966 M * fs ok dnetc --bench started in both the host and the vserver
1159912981 M * Bertl dnetc? what's that?
1159912995 M * fs something like boing, a silly tool sucking CPU
1159912996 M * ntrs_ daniel_hozac, how do I config cpusets with 211-rc2?
1159913000 M * fs a good test for this 
1159913002 M * daniel_hozac see doc/configuration.html
1159913008 M * ntrs_ ok
1159913012 M * daniel_hozac or vserver - build --help
1159913024 M * fs distributed.net client
1159913039 M * fs on vtop I have
1159913041 M * fs  3533 root      25   0  1224  492  420 R 50.3  0.0   0:37.66 dnetc              
1159913041 M * fs  3534 root      25   0   580  360  292 R 49.9  0.0   0:36.84 dnetc              
1159913053 M * fs both nice 0
1159913087 M * Bertl so they get the same priority
1159913091 M * fs yes
1159913099 M * fs I did not touch those values
1159913106 M * Bertl nevertheless one makes X jerky, the other not?
1159913116 M * fs exactly
1159913138 M * fs I tried renicing X to -5 and -10, it made no difference
1159913143 M * Bertl very interesting
1159913171 M * Bertl could you do me a favor and try with the following hog instead of yours
1159913177 M * fs yes
1159913182 M * Bertl http://vserver.13thfloor.at/Experimental/TOOLS/cpuhog-0.02.c
1159913246 M * Bertl is the hardcpu scheduler compiled into the kernel?
1159913266 M * Bertl maybe you could upload the .config to paste.linux-vserver.org?
1159913292 M * fs it's standard debian kernel 2.6.18-2
1159913312 M * fs # CONFIG_VSERVER_HARDCPU is not set
1159913315 M * fs if you mean this
1159913378 M * Bertl ah, hmm, debian kernel, probably I'm asking a lot of you when I ask you to try with a mainline 2.6.18 kernel plus vserver patch?
1159913391 M * waldi it is mainline
1159913407 M * Bertl okay, so we have no debian specific patches?
1159913408 M * fs the debian kernel is pretty close to vanilla, it contains only some added pci ids and the areca driver from 2.6.19-git
1159913421 M * Bertl okay, that's perfectly fine then
1159913431 M * fs no chimera like the redhat one =)
1159913517 M * fs  3588 root      22   0  3744  400  312 R 59.0  0.0   0:03.01 a.out              
1159913517 M * fs  3589 root      20   0  3728  400  316 R 41.7  0.0   0:02.63 a.out              
1159913526 M * fs the cpuhog running
1159913533 M * Bertl and does it affect X?
1159913566 M * fs yes, but not that bad as dnetc
1159913579 M * Bertl but again, only inside a guest?
1159913609 M * fs if I stop it in the host the mouse literally freezes
1159913648 M * daniel_hozac so while you're running the one in the host, the mouse is fine?
1159913652 M * fs if it runs in the host, everything is fine - in both, it's sluggish, in the vserver alone it freezes
1159913678 M * micah daniel_hozac: when do you expect to have 0.30.211 released?
1159913710 M * Bertl fs: okay, let's try to run it like this (on the host)
1159913726 M * Bertl chcontext --xid 666 -- ./a.out
1159913737 M * micah CONTEXT OF THE BEAST!
1159913740 M * daniel_hozac micah: ASAP, all that's left is basically some secure-mount fixes (now that internal handling is used, it showed some problems), mugwump telling me fai works, and a review of the differences.
1159913743 A * micah makes a symbol of the cross
1159913778 M * Bertl micah: you know what 668 is?
1159913788 M * micah daniel_hozac: great, I was asking because debian etch freeze is october 18th, so it would be good to have that in by then if possible 
1159913796 M * fs Bertl: New security context is 666 
1159913800 M * fs and mouse freezes
1159913802 M * daniel_hozac micah: yeah, h01ger told me about that the other day :)
1159913807 M * fs jumps
1159913810 M * micah Bertl: no, but I'm expecting it to be amusing :)
1159913813 M * Bertl fs: hmm, excellent ...
1159913839 M * Bertl micah: well, the neighbour of the beast, of course :)
1159913878 M * micah lol
1159913900 M * Bertl fs: how much time/efford would it be to recompile the kernel with hardcpu and idle time enabled?
1159913968 M * h01ger Bertl, 667 is the neighbour of the beast :)
1159913981 M * daniel_hozac h01ger: but on the other side of the street! :)
1159914007 M * Nei http://img96.imageshack.us/my.php?image=miiup6.png :>
1159914012 M * micah haha
1159914035 M * mugwump LOL e-mail from VMWare talking about their leading the transition to 64-bit computing
1159914069 M * Bertl ah, they finally got it working? :)
1159914073 M * mugwump It's like, I've been doing virtualisation on amd64 for what... about 3 years now
1159914119 M * _are_ mugwump: yes, but real 64-bit, they talk about transition and probably want to say 'emulating 32-bit only hardware on 64 bit" ;)
1159914157 M * ntrs_ daniel_hozac, I am reading the doc about cpusets, but I need some example. What is a good sane collection of settings for thost 6-7 files?
1159914171 M * daniel_hozac ntrs_: you only have to set name.
1159914180 M * ntrs_ just name?
1159914193 M * ntrs_ but I want to give one guest access to only one cpu.
1159914194 M * daniel_hozac hmm, seems you need mems too, maybe.
1159914203 M * ntrs_ what are mems?
1159914204 M * daniel_hozac right, but name is the only one required.
1159914213 M * daniel_hozac what memory nodes are available to the cpuset.
1159914220 M * ntrs_ what are memory nodes?
1159914233 M * daniel_hozac that i'm not entirely sure of :)
1159914243 M * ntrs_ so, how do I know what to set it to?
1159914254 M * daniel_hozac i believe normal boxes only have one memory node.
1159914277 M * ntrs_ so, do I set it to "1", "one", "blah", what?
1159914280 M * fs Bertl: I'll recompile it with those settings tomorrow at work and give it a try
1159914295 M * daniel_hozac ntrs_: i have it set to 0... i don't know if it's required though.
1159914306 M * daniel_hozac ntrs_: and note that you have to mount the filesystem on the host.
1159914316 M * daniel_hozac ntrs_: cpuset                  /dev/cpuset             cpuset  defaults        0 0
1159914335 M * ntrs_ in fstab on the host?
1159914338 M * daniel_hozac right.
1159914346 M * ntrs_ the fstab file for the guest?
1159914384 J * Aiken ~james@tooax7-094.dialup.optusnet.com.au
1159914384 M * daniel_hozac you don't have to change that.
1159914410 M * ntrs_ hmm, ok
1159914435 M * ntrs_ ok, how do I set a guest to have access to cpu1 only?
1159914438 M * fs thanks for the help, gotta go =) n8
1159914448 M * daniel_hozac ntrs_: 1 in cpus
1159914477 M * ntrs_ ok, daniel_hozac let me try this
1159914480 M * ntrs_ I assume I
1159914492 M * ntrs_ I assume I'll have to restart the guest for the changes to take effect.
1159914507 M * daniel_hozac yep.
1159914652 J * derjohn2 ~aj@dslb-084-059-025-169.pools.arcor-ip.net
1159914659 M * Bertl welcome derjohn2!
1159914682 M * Bertl fs: okay, tx, cya!
1159914743 M * derjohn2 Hi Bertl ;)
1159914745 M * ntrs_ derjohn2, daniel_hozac mentioned that you were able to reproduce the mysql hanging with 2.1.1-rc35. Can you give me some more information about that? I have the same problem it seems.
1159914802 M * ntrs_ daniel_hozac, the name of the cpuset should be unique, and could be the same as the guest's name, right?
1159914817 M * daniel_hozac right.
1159914832 M * derjohn2 ntrs_, Hi! Well, reproduce might be the wrong term. I notices mysql hanging in a form that you are able to login viy mysql DBname, I was able to type the query, but did not receive an answer.
1159914853 M * ntrs_ that is exactly what is happening in my case too.
1159914860 M * ntrs_ Bertl, it looks like the same issue.
1159914868 M * derjohn2 ntrs_, in other cases the mysql shell did hang in a way I had to close the Kconsole window and log in to the server again.
1159914879 M * ntrs_ derjohn2, this happens after about a day or so?
1159914884 M * derjohn2 ntrs_, but this was on devel rc31 !!!
1159914892 M * derjohn2 I thought it was fixed in rc35 ?
1159914897 M * Bertl derjohn2: okay, did you get around testing with rc35?
1159914900 M * ntrs_ derjohn2, you are not having the issue with rc35?
1159914916 M * derjohn2 ntrs_, a about 500 MB Database, only slightly loaded. I somestimes happens only after 20 days.
1159914963 M * derjohn2 Bertl, ntrs_ : I didnt upgrade the machine in question. other machines with different rc's are just fine with mysql quests.
1159914996 M * derjohn2 But I can lookup version details of the hanging machine. let me look, maybe we find similarties?
1159915010 J * are|lunch are@dslb-084-057-179-145.pools.arcor-ip.net
1159915059 M * daniel_hozac Bertl: oh, was this the problem the __enter/leave_vx_admin changes in __do_IRQ were about?
1159915067 M * Bertl derjohn2, ntrs_: yes, please also verify nptl vs. linux threading
1159915088 M * Bertl daniel_hozac: well, we thought so, but apparently the issue is back with rc35
1159915108 M * daniel_hozac right, i remember now.
1159915128 M * Bertl daniel_hozac: I'm not even sure the problem _is_ vserver related at all, could as well be some strange mysql behaviour which is exposed in a vserver env
1159915139 M * daniel_hozac yeah, i agree.
1159915177 M * derjohn2 2.6.16.14-p3-vs2.1.1-rc19-squash-drbd, AMD Athlon(tm) 64 X2 Dual Core Processor 3800+,3233 RAM, Debian GNU/Linux testing/unstable \n \ guest, mysql  4.1.15-Debian_1-log
1159915205 M * derjohn2 nptl vs.  linux threading : how to look that up ????
1159915238 M * ntrs_ derjohn2, yes good question.
1159915241 M * derjohn2 ntrs_, did you already create a monitor for auto-restart of the mysql ?
1159915255 M * ntrs_ derjohn2, no, how do I do that?
1159915266 M * ntrs_ did you create one that you'd like to share?
1159915280 M * derjohn2 I do something with monit, but I dont know if it works, since the errro didnt occur again
1159915308 M * derjohn2 * *     * * *   root     echo "select *,now() from CMS_SYSTEMID;" | mysql xxxxx > /etc/monit/qimondadb-output
1159915316 M * derjohn2 then I check the filedate ;)
1159915345 M * derjohn2 check file mysqltestoutput with path /etc/monit/qimondadb-output if timestamp > 2 minutes then alert ....
1159915355 M * derjohn2 *ouch*
1159915406 M * derjohn2 i didnt set a restart yet
1159915431 Q * _are_ Ping timeout: 480 seconds
1159915441 M * derjohn2 I created the script 2006-09-15,l since then the db runs.
1159915470 M * ntrs_ daniel_hozac, does the values in cpus start from 0 or 1?
1159915480 M * daniel_hozac ntrs_: 0, same as for mems.
1159915502 M * ntrs_ derjohn2, that is a good uptime. My guest has a hanging mysql within a day
1159915507 M * derjohn2 I dont really know if the "filewriter" would modify the file timestamp. if there is no query output ...
1159915513 M * ntrs_ daniel_hozac, I thought I don't need mems, or do 
1159915516 M * ntrs_ i?
1159915525 M * daniel_hozac ntrs_: i don't think so... i don't know why i have it.
1159915529 M * derjohn2 maybe you want to rest the monitor? I can mail you all files I use ...
1159915542 M * derjohn2 *test :)(
1159915586 M * ntrs_ derjohn2, I'd rather see the problem fixed with no need for monitoring and restarting, we'll see
1159915642 M * derjohn2 ntrs_, the problem is there since rc19, reported by you since rc29 or so. I would sleep better, if the machine would pull itself up ATM ...
1159915663 M * derjohn2 BTW: /etc/init.d/mysql restart helps to resurrect the DB 
1159915674 M * ntrs_ daniel_hozac, how do I mount cpuset properly on the command line?
1159915682 M * daniel_hozac mount -t cpuset cpuset /dev/cpuset
1159915688 M * ntrs_ derjohn2, yes that is correct.
1159915697 M * daniel_hozac (after creating the directory)
1159915716 M * derjohn2 why dont others experience the problem? ntrs_ are you on dualcore athlon, too?
1159915743 M * ntrs_ derjohn2, no it happens on different machines, this particular one is a dual PIII
1159915748 M * derjohn2 again: nptl vs.  linux threading : how to look that up ????
1159915761 M * ntrs_ daniel_hozac, # mount -t cpuset cpuset /dev/cpuset
1159915761 M * ntrs_ mount: fs type cpuset not supported by kernel
1159915764 M * derjohn2 ntrs_, but it'S dual ... 
1159915773 M * daniel_hozac ntrs_: that'll be a problem then :)
1159915778 M * ntrs_ derjohn2, yes it is
1159915789 M * Bertl derjohn2: I don't know which is which, the question is more, does ps -m show threads or do you see the threads without -m
1159915791 M * daniel_hozac ntrs_: CONFIG_CPUSETS
1159915794 M * ntrs_ 2.6.17.13-RH211rc35
1159915796 M * derjohn2 maybe the timing problem only occurrs on multi cpu?
1159915807 M * ntrs_ derjohn, could be
1159915815 M * ntrs_ that's why I want to try the cpusets.
1159915823 M * Bertl derjohn2: could as well be a multi cpu threading issue with mysql
1159915920 M * derjohn2 Bertl, silly question: how would it look like seeing threads? 
1159915943 M * derjohn2 Bertl, very unlikely .. ntrs_ which mysql version?
1159915993 M * ntrs_ derjohn, I've seen it with 4.1.x and 5.0.x
1159916051 M * ntrs_ I always miss some kernel option. I'll rebuild now with cpusets. I thought that was part of the vserver patch and there was no option required.,
1159916078 M * Bertl derjohn2: http://paste.linux-vserver.org/429
1159916095 M * derjohn2 ntrs_, I also raised the keybuffer im mysql a lot ...
1159916098 M * ntrs_ Do I need some other kernel options to be able to use all the new vserver options in devel?
1159916132 M * ntrs_ derjohn2, unfortunately, I can't do that. The customers are the only ones that can touch their mysql options
1159916143 M * daniel_hozac ntrs_: cpusets is an entirely different kernel feature, we just use it for limiting in the utils.
1159916178 M * ntrs_ Do I need some other kernel options to be able to use all the new vserver options in devel?
1159916183 M * derjohn2 ntrs_, so they use the deafults?
1159916209 M * ntrs_ derjohn2, mostly, I assume.
1159916274 M * daniel_hozac ntrs_: that depends on the options you have enabled :) but i think cpusets is the only conditionally built feature the utils use.
1159916288 M * ntrs_ ok
1159916333 M * derjohn2 Bertl, I see "mysql        -  0.0    -      -     - -        Sl   Sep15   0:00 -" ...means?
1159916360 Q * bubulak Ping timeout: 480 seconds
1159916468 J * bubulak ~bubulak@whisky.pendo.sk
1159916508 M * Bertl derjohn2: are there more mysql entries with -m than without (ignoring the main entry)
1159916709 M * derjohn2 Bertl, http://paste.linux-vserver.org/430
1159916730 M * derjohn2 only one "mysql", all others are therads?
1159916880 M * Bertl okay, so that uses native linux threads
1159916957 M * ntrs_ Bertl, that seems to be more stable than the other one
1159917726 M * ntrs_ how can I verify that I am using cpusets now and how can I verify that a guest is really limited to cpu0
1159917728 M * ntrs_ ?
1159917787 M * daniel_hozac try two cpuhog and see how much CPU they use?
1159917804 M * daniel_hozac cat /proc/self/status should show a cpuset line, IIRC.
1159917813 M * ntrs_ daniel_hozac, do I need to recreate the directory on every reboot?
1159917832 M * daniel_hozac it depends on your host's OS, if it's using a tmpfs /dev mount, you will.
1159917864 M * ntrs_ I think it's using udev
1159917939 M * daniel_hozac mkdir -p /etc/udev/devices/cpuset/cpuset
1159918358 M * ntrs_ # cat /proc/virtual/111/sched
1159918358 M * ntrs_ FillRate:             10,1
1159918358 M * ntrs_ Interval:            100,8
1159918358 M * ntrs_ TokensMin:            10
1159918358 M * ntrs_ TokensMax:           400
1159918360 M * ntrs_ PrioBias:              0
1159918361 M * ntrs_ VaVaVoom:              7
1159918363 M * ntrs_ cpu 0: 10072 1532 13043 6900 0 R- 27 10 400 10/100 1/8
1159918365 M * ntrs_ cpu 1: 9529 1402 8630 7040 0 R- 74 10 400 10/100 1/8
1159918367 M * ntrs_ it looks like it's still using both cpus
1159918399 M * daniel_hozac hmm, did you restart it? what settings did you use?
1159918411 M * daniel_hozac what does cat /dev/cpuset/<name>/tasks say?
1159918448 M * ntrs_ daniel_hozac, yes I restarted the guest, name (somename), cpus (0)
1159918488 M * ntrs_ there is no "somename" in /dev/cpuset
1159918499 M * ntrs_ there is /dev/cpuset/tasks
1159918512 M * daniel_hozac hmm, are you sure you restarted the guest?
1159918537 M * ntrs_ yes, I am sure. I will restart it again just in case.
1159918543 M * daniel_hozac and where did you place those files? /etc/vservers/<name>/cpuset, right?
1159918557 M * derjohn2 ntrs_, funny, I upgraded the mysql in etch and found the following differnce in the my.cnf: # expire-logs-days      = 20    FIXME: Currently this setting causes mysqld to
1159918557 M * derjohn2 #                                      crash (debian#368547 mysql#17733)
1159918577 M * daniel_hozac derjohn2: haha!
1159918614 M * derjohn2 well, I my old conf  expire-logs-days was enabled.
1159918626 M * ntrs_ derjohn2, hmm, let me check
1159918721 M * ntrs_ derjohn2, no, there is no such setting here
1159918735 M * ntrs_ daniel_hozac, still no "somename" in /dev/cpuset
1159918767 M * daniel_hozac ntrs_: are you sure you're using a recent util-vserver release?
1159918780 M * daniel_hozac grep CPUSET /usr/lib*/util-vserver/vserver.start
1159918781 M * ntrs_ 211-rc2
1159918786 M * ntrs_ I just rebuilt it
1159918804 M * ntrs_ # grep CPUSET /usr/lib*/util-vserver/vserver.start
1159918804 M * ntrs_ addtoCPUSET  "$VSERVER_DIR"
1159918805 M * derjohn2 ntrs_, I cannot tell, if that is respnsible for the problem - I have a different one but maybe related one on a differnt host "mysql bbba" suddenly freezes.
1159918882 M * daniel_hozac ntrs_: _very_ strange...
1159918914 M * daniel_hozac ntrs_: and you used the correct name for the directory?
1159918950 M * ntrs_ for what directory, in /dev?
1159918987 M * daniel_hozac hmm, both, i suppose.
1159918989 M * ntrs_ # ls /dev/cpuset/
1159918990 M * ntrs_ cpu_exclusive  mem_exclusive   memory_pressure          memory_spread_page  mems               tasks
1159918990 M * ntrs_ cpus           memory_migrate  memory_pressure_enabled  memory_spread_slab  notify_on_release
1159918996 M * daniel_hozac it's cpuset in both instances, anyway.
1159919037 M * ntrs_ # ls /etc/vservers/casa82it/cpuset/
1159919038 M * ntrs_ cpus  name
1159919176 M * daniel_hozac hmm, i must be missing something.
1159919181 M * ntrs_ I guess
1159919189 M * daniel_hozac what happens if you run mkdir /dev/cpuset/casa82it?
1159919248 M * daniel_hozac do you by any chance have namespace cleanup enabled?
1159919435 M * ntrs_ no, I don't have namespace cleanup
1159919485 M * ntrs_ I just created that dir and am restarting the guest
1159919515 M * daniel_hozac so it was created?
1159919549 M * ntrs_ I created it manually, after the restart of the guest the dir is gone
1159919559 M * daniel_hozac do you get any errors during the start?
1159919602 M * daniel_hozac that's a good sign though, because it means the right code is getting executed, and it's just failing.
1159919678 M * ntrs_ no, I don't see any errors on start
1159919691 M * daniel_hozac do you see anything on start?
1159919723 M * daniel_hozac because the _only_ reason for the directory to be removed is if it failed.
1159919728 M * ntrs_ no, nothing
1159919731 M * daniel_hozac in which case you should be getting a warning.
1159919739 M * daniel_hozac hmm, do you have apps/init/tty set?
1159919750 M * ntrs_ oh, wait
1159919769 M * ntrs_ I have this on start
1159919770 M * ntrs_ WARNING: Failed to create or CPUSET "casa82it" does not exist! Not using it!
1159919880 M * ntrs_ aren't we able to use separate scheduling per CPU?
1159919890 M * daniel_hozac hmm?
1159919919 M * ntrs_ set separate sched parameters per cpu?
1159919932 N * hardwire hardwire-afk
1159919934 M * daniel_hozac oh yes, but not with util-vserver.
1159919935 N * hardwire-afk hardwire
1159919944 M * daniel_hozac don't you get anything before that?
1159919945 M * ntrs_ with what then?
1159919956 M * daniel_hozac vcmd should be able to.
1159919965 M * ntrs_ no there is nothing before that