1133829294 Q * dlippolt Quit: Leaving
1133829635 N * Bertl_oO Bertl
1133829641 M * Bertl back now ...
1133829762 M * matti :)
1133829775 A * matti hugs Bertl :)
1133829781 M * matti Well, goodnight all :)
1133829790 M * Bertl sweet dreams ...
1133829793 M * matti Bertl: Remember, you need sleep too ;)
1133829794 M * matti Thanks.
1133829805 M * [MUPPETS]Gonzo USA is good for you, bertl, just the same day/night scheme you are used to have at home (but still same timezone) :)
1133829825 M * Bertl [MUPPETS]Gonzo: yes, was quite easy to adapt
1133829838 M * matti :-P
1133829854 M * Bertl [MUPPETS]Gonzo: but it's getting harder and harder as my timezone slowly starts to shift :)
1133829891 M * [MUPPETS]Gonzo ah, time to get back. AFAIK in one week?
1133829937 M * Bertl derjohn: you mean linux-vserver ontop of xen 3.0?
1133830033 M * Bertl derjohn: do you have an url at hand where to get the xen modifications (against a moderately recent vanilla kernel)?
1133830036 M * derjohn nom i thought some xens in a linux-vserver ... <g> ... yes 
1133830059 M * derjohn Bertl, nope. I am strictly vserverish at the time.
1133830077 M * [MUPPETS]Gonzo  Debian -- kernel-patch-xen - (http://packages.debian.org/unstable/misc/kernel-patch-xen)
1133830079 M * derjohn but I am thinking about ... XEN 
1133830099 M * [MUPPETS]Gonzo http://packages.debian.org/unstable/source/xen
1133830116 M * derjohn Bertl, I hope that Mirgarion of 'old hosts' could be easier with XEN (no init fiddeling etc) 
1133830164 M * derjohn Bertl, one Q left: what is new with RC9 ?
1133830178 M * [MUPPETS]Gonzo maybe it would be better to throw away theese old hosts, john...
1133830233 M * derjohn [MUPPETS] Gonzo, yes, it would. 
1133830356 M * derjohn [MUPPETS]Gonzo, you should change your nick to [LINUX-VSERVER]Gonzo 
1133830373 M * matti :P
1133830382 A * matti just saw muppets from space :)
1133830382 M * matti ;p
1133830389 M * [MUPPETS]Gonzo cool film
1133830393 M * matti Yeah :)
1133830395 M * [MUPPETS]Gonzo now you know, where i come from
1133830396 M * Bertl derjohn: hmm, what is supposed to be with rc9?
1133830396 M * [MUPPETS]Gonzo *g*
1133830405 M * matti Hehehe.
1133830431 M * derjohn Bertl, just wanted to know what the issue with rc8 was (just finished my rc8 testing and began rollout ....)
1133830442 M * anonc Bertl: re the COW fixes in rc9 - successfully linked and launched 253 debian sid vservers on the one host - no errors and COW seems to function as expected (though it seemed to work whether or not I set the iunlink flag...)
1133830491 M * derjohn anonc, COW? 
1133830499 M * anonc copy on write
1133830523 M * anonc there was an issue with hardlinking devices that caused panics
1133830553 M * derjohn anonc, devices only? not with "real-file" hardlinks?
1133830561 M * derjohn anonc, anything else ?
1133830620 M * anonc derjohn, not as far as i know - I believe it now checks the type of file hardlinked and only operates on those that are safe
1133830667 M * derjohn anonc, OK. But then I will skip all rc8 stuff and build a RC9 release of my debian kernel pack.
1133830679 M * derjohn anonc, THX for the info.
1133830699 M * Bertl derjohn: ah, well, did you check the differences?
1133830704 M * anonc derjohn, THX to betrl for the fixes
1133830749 M * derjohn Bertl, no - i am a "lame c reader". The rcs dont have a changelog (?)
1133831003 M * Bertl nope sorry, ext3 has no changelogs yet :)
1133831003 Q * eyck Read error: Connection reset by peer
1133831009 M * [MUPPETS]Gonzo n8 derjohn
1133831030 M * Bertl but let me have a look at the changes for ya :)
1133831043 M * derjohn I think I have to .zZ a little
1133831055 M * derjohn Bertl, hm, you dont know them by hard?
1133831084 M * Bertl I know and I forget them when released, as they are in the code anyway
1133831110 M * Bertl here are the changes:
1133831118 M * derjohn Bertl, interesting philosophy :) To b short: I will switch to rc9.
1133831126 M * micah derjohn: ah, you have not responded to my email :)
1133831130 M * Bertl - procfs bugfix (possible oops)
1133831141 M * derjohn micah, yes - i waited for rc9 ;)
1133831160 M * Bertl - reiserfs dlimit block allocation fix
1133831161 M * micah derjohn: I was just updating the package to have -rc9 now
1133831190 M * Bertl - various space cleanups
1133831191 M * derjohn micah, besides that I wanted a symmetric timespan between all out mail ;)
1133831197 M * micah hehe
1133831216 M * Bertl - fix for the COW oops (on non regular files)
1133831218 M * derjohn Bertl, OMG there was a bug with my beloved reiser?
1133831225 Q * monrad Quit: leaving
1133831236 M * [MUPPETS]Gonzo reiser sucks :-p
1133831247 M * Bertl - dynamic xid option (with new default, be careful)
1133831272 M * Bertl that's it :)
1133831281 M * ag- [MUPPETS]Gonzo: very clever comment...
1133831286 M * Bertl derjohn: yeah, your beloved reiserfs didn't account properly :)
1133831303 M * [MUPPETS]Gonzo ag-: sorry to make an insider-joke public
1133831305 M * derjohn Bertl, sounds like a woman
1133831309 M * micah derjohn: do you have the 64 IP/guest patch along with -rc9? I can include that in this upload
1133831349 M * derjohn micah, I got that file as an old delta. but you would have to patch util-vserver anyway to makr it work.
1133831377 M * micah derjohn: I can get that patch into the debian util-vserver, I'm starting to have influence on Ola
1133831385 M * derjohn micah, I dont know if it could do harm if you use unpatched utils
1133831445 M * derjohn micah, the patch was an issue because it slows down networking (linear lookup). 
1133831470 M * ag- micah: NMU power!
1133831491 M * micah derjohn: you mean the patch that enables the 64 IPs causes a slowdown in networking?
1133831508 M * micah ag-: hehe, I could... but I am trying to be nice and get co-maintainership
1133831613 M * derjohn micah, thats what they told me. I don't think it's much so. But for some reason all exceopt me are lucky with 16 IPs.
1133831675 M * derjohn micah, mail is out. The patch does not apply clean, but the failed hunk is useless in 2.6.14/rc8, I think in rc9 too.
1133831718 M * micah derjohn: hmm, slowing down networking sounds like a regression
1133831801 M * derjohn micah, yes. I would be careful with the patch. It you like I create a clean diff we^W you could pack into /usr/share/doc/util-foo 
1133831846 M * derjohn micah, or into kernel-patch-foo ....
1133831875 M * derjohn (as long as you aren't co-maintainer)
1133831879 M * micah hmm
1133831910 Q * matti Ping timeout: 480 seconds
1133832079 M * micah derjohn: NB_IPV4ROOT is in 4 files total, include/linux/vserver/legacy.h and network.h AND kernel/legacy.h and network.h
1133832096 M * derjohn micah, I dont want to be inpolite but I have to leave now (-> sleep). 
1133832114 M * micah derjohn: its ok
1133832121 M * micah good night!
1133832160 M * derjohn micah, hm, 4 files? seen from diff or cirrent kernel source?
1133832286 M * derjohn micah, I had one hunk fail AFAIR - the define at the particular file was replaced in newer RCs...
1133832348 M * derjohn micah, we stay in contact .... if you have tetsing package, please mail be the link! (people.debain?))
1133832355 M * derjohn over and out!
1133832423 M * micah yeah I dont think its a good idea to include this patch in the debian packages, slowing down networking in an unstable way seems like a worse option than having people not use so many IPs per guest :)
1133832668 A * micah reads past IRC conversations about the issue
1133832755 M * Bertl micah, derjohn, whoever: bring me at least 4 folks who have a good reason to 'require' more than 16 IPs and we do a proper solution for the legacy code
1133832918 M * micah i haven't found a reason myself yet, I have 5 IPs in a guest because of SSL certificates, but thats not even close to 16 yet
1133832983 M * micah Bertl: I've seen four... shuri, derjohn, infowolfe and nokoya in the IRC logs who were asking for it. Now I can't tell you if they have a "good reason to 'require'" :)
1133833011 M * micah i only know that because I just read all the occurances of the logs that referenced NB_IPV4ROOT to get an idea of what the issue was
1133833013 M * Bertl yeah, that's the problem, most folks (out of the blue) would like to have it
1133833033 M * micah none of them have said *why* they need/want it
1133833125 M * micah i dont plan on putting it into the debian kernel patch
1133833587 M * Bertl basically we could do some kind of hash lookup and/or support network ranges
1133833604 M * Bertl but I do not want to put my time into that when there is no real need
1133833614 M * Bertl (especially as it would be legacy only)
1133833635 M * [MUPPETS]Gonzo derjohn had a real need, but I don't think the setup is quiet usual
1133833765 M * micah oh, the IP limit is a legacy issue? I didn't realize that
1133833830 M * Bertl yes, it is even worse, that's why the tools require patching
1133833849 M * Bertl even worse means: the number is hardcoded in the userspace-kernel interface
1133833869 M * micah oh legacy mode, when will you go away? :)
1133833953 M * Bertl asap, Hollow's tools should be there (at least pretty soon), and I hope enrico will pick it up too
1133833969 M * micah I am looking forward to his tools
1133836817 J * mep_ mep@p5091A5F8.dip0.t-ipconnect.de
1133836922 M * Bertl welcome mep_!
1133837055 J * monrad ~mikkel@213083190131.sonofon.dk
1133837098 M * Bertl evening monrad!
1133837165 Q * monrad Quit: 
1133837173 M * [MUPPETS]Gonzo any chance to define the directory where the "to be build" vserver should be placed? I found --rootdir, but this creates a directory with the vserver name inside the rootdir
1133837225 M * Bertl a) what tool version, b) what configuration?
1133837252 M * [MUPPETS]Gonzo vserver 0.30.209
1133837258 Q * mep__ Ping timeout: 480 seconds
1133837270 M * [MUPPETS]Gonzo please specify question b :)
1133837292 M * Bertl vserver-info - SYSINFO (please upload somewhere)
1133837322 J * matti matti@linux.gentoo.pl
1133837350 M * [MUPPETS]Gonzo http://pastebin.com/450428
1133837353 M * Bertl hmm, wb matti!
1133837367 M * [MUPPETS]Gonzo debian sid, fresh install
1133837395 M * Bertl well, guests will be located in /var/lib/vservsers (as specified)
1133837406 M * matti Bertl: :)
1133837445 M * [MUPPETS]Gonzo yes. I tried to specify the directory, in my example: vserver vorlage-inside build -m debootstrap --hostname vorlage-inside --rootdir /vservers/vorlage/BACKUP/ [etc]
1133837459 J * monrad ~mikkel@213083190131.sonofon.dk
1133837462 M * [MUPPETS]Gonzo but it created /vservers/vorlage/BACKUP/vorlage-inside/
1133837477 M * [MUPPETS]Gonzo but I want it to install into /vservers/vorlage/BACKUP/ directly
1133837492 M * [MUPPETS]Gonzo is this possible?
1133837498 M * Bertl ahem, then you should call your guest BACKUP
1133837523 M * Bertl vserver BACKUP build -m debootstrap --hostname vorlage-inside --rootdir /vservers/vorlage ...
1133837548 M * [MUPPETS]Gonzo this is not possible, as I want to have ... mhm, let my write a short description, one sec
1133837575 M * Bertl I don't think the tools ever considered that possibility
1133837598 M * Bertl I have to say that I do not see a good reason for it either ...
1133837796 M * [MUPPETS]Gonzo mind if I PM you the description?
1133837822 M * Bertl well, no, but I think it would be on-topic here ..
1133837834 M * [MUPPETS]Gonzo http://pastebin.com/450436
1133837854 M * [MUPPETS]Gonzo not easy to describe my minds in english, sorry
1133837908 M * Bertl well, looks like a 'simple' rsync failover setup
1133837928 M * Bertl now my questions are:
1133837989 M * Bertl - why not call both, the original and the backup(s) 'gonzo'?
1133838009 M * Bertl - what are 'external' hosts? real machines?
1133838047 M * Bertl would make more sense to me to have, let's say
1133838059 M * [MUPPETS]Gonzo because they have different IP setups. the host being up in "normal times" only uses an internal ip, the "emergency host" is supposed to fire up the real "external" ip and yes, external host: real machine
1133838061 M * Bertl a directory /vservers and another one called /vbackups
1133838092 M * Bertl and have two configs for them, called gonzo and gonto-failover or so
1133838099 M * Bertl *gonzo
1133838115 M * Bertl or, alternatively, have them just named differently
1133838141 M * Bertl (should be no big deal to rsync gonzo to gonzo-backup, no?)
1133838158 M * [MUPPETS]Gonzo the reason for using two vservers is simple: If I want to make a complete rsync backup including all file-rights, I have to be root. Several external machines shall rsync-backup to this host, so each of them has to have a root-account. For security, they can not have root on the same machine 
1133838194 M * [MUPPETS]Gonzo thats why i thought about setting up a vserver for each backup, containing just rsync and ssh
1133838214 M * [MUPPETS]Gonzo and inside that directory I wanted to have the rsynced-backup
1133838222 M * [MUPPETS]Gonzo which I want to fire up if something happens
1133838222 M * Bertl well, what about this setup then
1133838313 M * Bertl - have a single (or if you really prefer multiple) guest(s) called 'backup' (or backup-*) which just runs a minimalized (probably unified) ssh/rsync setup, and rsync the guests into this special guest(s) at /vservers/backup/backup
1133838335 M * Bertl - setup the failover with a root path of /vservers/backup/backup
1133838355 M * Bertl - have the failover guests start within /vservers/backup/backup/gonzo (e.g.)
1133838363 M * anonc I've been toying with what writeable snapshots would offer for vservers and backups/testing...I *think* it ends up giving you block level copy-on-write - but I doubt its any good for running long term.
1133838401 M * Bertl well, it's what you get with UML ... better than nothing, much worse than unification
1133838475 M * [MUPPETS]Gonzo so I would have vserver1 start in /vservers/backup-gonzo/       and vserver2 in /vservers/backup-gonzo/backup/gonzo/?
1133838503 M * Bertl yes, while vserver1 would be the ssh/rsync guest and vserver2 the failover
1133838525 M * Bertl (which would probably stop vserver1 before starting up)
1133838614 M * [MUPPETS]Gonzo but then the external host could not use vserver1:/backup as destination dir but would have to know that it is called vserver1:/backup/gonzo - which I try to avoid
1133838637 M * Bertl well, that's pretty easy, thanks to symlinks, no?
1133838661 M * Bertl if you don't like them, you could also --bind mount the dir to /BACKUP
1133838726 M * [MUPPETS]Gonzo that's an idea
1133838729 M * [MUPPETS]Gonzo damn
1133838734 M * [MUPPETS]Gonzo I'll try it now
1133838745 M * Bertl excellent, let us know how it works out
1133838760 M * Bertl btw, I'd suggest to use ssh keying and a single backup guest
1133838808 M * [MUPPETS]Gonzo I'll use ssh keys. The problem with only one guest is, that you could see all files of the others backups
1133838810 M * [MUPPETS]Gonzo which is not wanted
1133838843 M * Bertl if you specify a certain command for the rsync, you should be able to prevent that
1133838860 M * Bertl you do not need to allow ssh access to the backup machine beyond the rsync
1133838964 M * [MUPPETS]Gonzo mhm, by using own vservers for each backuphost, I can control disk/cpu quota in a better way
1133838973 M * [MUPPETS]Gonzo and can move it to another machine easily
1133838979 M * [MUPPETS]Gonzo so I would prefer that way
1133839002 M * Bertl k, no probelm with that :)
1133839083 Q * flock Ping timeout: 480 seconds
1133839235 J * ryker ~ryker@c-67-176-240-74.hsd1.in.comcast.net
1133839264 M * Bertl welcome ryker!
1133840072 M * [MUPPETS]Gonzo waiting for my test-rsync to finish...
1133840692 J * stefani ~stefani@c-24-19-46-211.hsd1.wa.comcast.net
1133840697 M * [MUPPETS]Gonzo looks like it kinda worked :)
1133840736 M * [MUPPETS]Gonzo need some fine tuning and hard-linked rsync backups, but so far it is very fine. I did not think about symlinks being possible
1133840741 M * [MUPPETS]Gonzo thanks a lot, bertl :)
1133840752 M * Bertl you're welcome!
1133840807 M * [MUPPETS]Gonzo I'll go to bed now, 4:47 here :)
1133840824 M * Bertl k, good night then :)
1133841281 J * _Medivh ck@paradise.by.the.dashboardlight.de
1133841282 Q * Medivh Read error: Connection reset by peer
1133842055 Q * Johnnie Remote host closed the connection
1133842117 J * Johnnie ~john@acs-24-154-53-16.zoominternet.net
1133842242 Q * Johnnie Remote host closed the connection
1133842564 J * Johnnie ~john@acs-24-154-53-16.zoominternet.net
1133844414 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il
1133845517 J * Aiken_ ~james@tooax7-225.dialup.optusnet.com.au
1133845597 M * micah http://lwn.net/Articles/162684/
1133845714 M * Bertl tx
1133845845 Q * Aiken Ping timeout: 480 seconds
1133846258 Q * ^Cist Quit: Leaving
1133847474 J * balbir ~balbir@59.145.136.1
1133847493 M * Bertl welcome balbir!
1133848121 M * Bertl okay, off to bed now .. back tomorrow!
1133848127 N * Bertl Bertl_zZ
1133848242 Q * balbir Quit: Leaving
1133848371 J * balbir ~balbir@59.145.136.1
1133850123 P * stefani parting (is such sweet sorrow)
1133850733 J * eyck eyck@81.219.64.71
1133850921 J * menomc ~amery@200.75.27.60
1133850934 J * Larsioa ~larspower@you.are.the.lamer.la
1133850935 M * Larsioa Hello
1133850989 M * Larsioa have anybody seen connection drops in vserver? I have a client thats hosting Ventrilo (Voice server) he tells me that every 30 mins or soo they are getting disconnected. If i host this ventrilo server outside the vserver (on the main) there are no problems.
1133851023 Q * Johnnie Remote host closed the connection
1133851031 Q * mnemoc Ping timeout: 480 seconds
1133851031 N * menomc mnemoc
1133852574 J * supastuff ~user@83.149.117.143
1133852576 M * supastuff hello
1133854069 Q * ryker Quit: Leaving
1133857288 J * nox_ ~nox@noxlux.de
1133857425 Q * nox Ping timeout: 480 seconds
1133858230 Q * shedi Quit: Leaving
1133859186 Q * _Medivh Quit: changing servers
1133859193 J * Medivh ck@paradise.by.the.dashboardlight.de
1133859243 Q * flock Ping timeout: 480 seconds
1133859293 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il
1133859824 M * sizo moin
1133860933 Q * flock Ping timeout: 480 seconds
1133860966 M * ag- Larsioa: proprietary software, uh?
1133861348 M * ag- Bertl_zZ: if openvz is pushed in mainstream, vserver won't be accepted as such, will it?
1133861466 M * SiD3WiNDR openvz?
1133861494 M * ag- SiD3WiNDR: url pasted from micah and http://openvz.org/
1133861549 A * SiD3WiNDR oogles
1133861568 M * ag- it kind of worry me, because to my mind, vserver is far better than ovz :)
1133861607 M * SiD3WiNDR hehe
1133861637 A * SiD3WiNDR supports vserver cause a) it Just Works(tm) b) support is very good and very fast c) it does everything I want it to do
1133861789 M * ag- SiD3WiNDR: exactly :) it's portable/ported on several hardware archs and the features ovz offers are planned for vserver
1133861810 M * SiD3WiNDR only thing I'd REALLY want now is v6support/ngnet
1133861956 M * ag- apprently, there's a possibility to add ipv6 without ngnet, but i read about issues with large buffers or such... i don't remember exactly :/
1133862019 J * shedi ~siggi@tolvudeild-197.lhi.is
1133862092 M * SiD3WiNDR :o
1133862964 J * robig ~Miranda@212.80.250.154
1133862973 M * robig hi
1133863019 Q * Ben81 Quit: Leaving
1133863090 M * daniel_hozac ag-: i think it's more of a "why work on the current when you can work on the next-gen?"
1133863108 J * Duckx ~duckx@195.75.27.158
1133863342 M * ag- daniel_hozac: for performance reasons, i think
1133863389 M * ag- daniel_hozac: Bertl_zZ wanted to keep the current network system for people desiring performances
1133863405 M * daniel_hozac ag-: there's that too, but the current structure could already hold 4 IPv6 addresses.
1133863419 M * daniel_hozac (IIRC, haven't really looked at IPv6...)
1133863421 M * ag- so obviously, adding ipv6 support to the current isn't time loose :P
1133863434 M * ag- s/loose/loss/
1133863463 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il
1133863480 M * daniel_hozac right, i personally like the current system.
1133863506 M * robig what is the current system?
1133863509 M * robig :)
1133863509 M * daniel_hozac (not that i don't like ngnet, i just have no reason to add overhead)
1133863533 M * daniel_hozac chbind
1133863707 M * ag- daniel_hozac: exactly :)
1133863715 M * ag- robig: current networking
1133863743 M * SiD3WiNDR so okayyyy.. who's gonna add v6 support! ;)
1133863789 M * daniel_hozac why don't you? ;)
1133863824 M * SiD3WiNDR I was afraid that question would come :>
1133863848 M * SiD3WiNDR if I can do it in java, php, perl or tcl.. :p
1133863885 M * daniel_hozac as C is far more straightforward than any of those, i don't see the problem :)
1133863942 A * SiD3WiNDR will severly break your kernel ;)
1133864082 M * ag- it's on my todo list under paid work :P
1133864101 M * ag- s/under/below/
1133864227 M * robig ag-:  is there a max open connections limit?
1133864434 M * daniel_hozac robig: in the current networking code? no. in the kernel? yes.
1133864568 M * robig and how do i find out the max connections of the kernel?
1133864581 M * daniel_hozac it's setable.
1133864587 M * daniel_hozac on a per-vserver basis as well.
1133864592 M * daniel_hozac openfd limit, IIRC.
1133864602 M * daniel_hozac (includes files as well)
1133864672 M * robig can u tell me where to get the currently open files/conn and the current max?
1133864750 M * daniel_hozac /proc/virtual/*/limit for per-vserver, /proc/sys/fs/file-max (i guess) for total.
1133864774 M * robig and youve written about different network systems. what systems are available? whats the one with the best performance?
1133864776 M * robig :)
1133864817 M * daniel_hozac the current system, and ngnet.
1133864843 M * daniel_hozac ngnet (will) virtualizes the network stack, so it adds some overhead.
1133864878 M * robig and the virtualisation is for what?
1133864921 Q * Aiken_ Quit: Leaving
1133864967 M * daniel_hozac i honestly don't know. someone with more ngnet insight will have to tell you :)
1133865073 M * robig ok. thank you :)
1133865081 M * robig back to the max connections
1133865113 M * robig cat /proc/virtual/x/limit shows a table of various numbers
1133865124 M * robig whats the meaning of that all? :)
1133865195 Q * flock Ping timeout: 480 seconds
1133865327 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il
1133865477 M * daniel_hozac http://linux-vserver.org/HowTo+Read+ProcFS
1133865502 Q * balbir Remote host closed the connection
1133865649 M * robig thanks. and what happend if e.g. max sock is reached?
1133865659 M * daniel_hozac creating new sockets is denied.
1133865681 M * robig and can i change the limits somewhere?
1133865705 M * daniel_hozac yes, see the flower page for /etc/vservers/<name>/rlimits
1133865734 M * daniel_hozac to change them on the fly, use vlimit.
1133865763 M * robig there's no rlimits file
1133865770 M * daniel_hozac it's a directory.
1133865779 M * daniel_hozac the flower page will tell you how to populate it.
1133865863 Q * flock Ping timeout: 480 seconds
1133865892 M * robig flower page?
1133865912 M * Loki|muh yeah
1133865915 M * Loki|muh nice page
1133865916 M * robig what do you mean with that?
1133865926 M * Loki|muh you should switch the css
1133865937 M * Loki|muh otherwise you will go mad ^^
1133865945 M * robig tell me the url please
1133865947 M * Loki|muh mom
1133865962 M * Loki|muh http://www.nongnu.org/util-vserver/doc/conf/configuration.html
1133865987 M * Loki|muh i like gras1
1133865988 M * Loki|muh ^^
1133866023 M * Loki|muh uh weedpage is new, didn't know that one
1133866039 M * Loki|muh nice one
1133866049 M * robig omg!
1133866064 M * robig where can i change the css?! tell me please!
1133866079 M * Loki|muh lol :)
1133866105 M * Loki|muh something within view
1133866110 M * SiD3WiNDR weedpage is like years old :p
1133866125 M * Loki|muh what?
1133866128 M * Loki|muh ups
1133866131 M * SiD3WiNDR "view - use style" in mozilla
1133866162 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il
1133866325 M * ag- robig: ngnet will allow you to define network interfaces inside a guest and to share them between guests and between guests and host
1133866372 M * ag- robig: you will be able to do networking stuff inside a guest, e.g. iptables, ebtables, and so on...
1133866398 M * ag- robig: and of course, managing routing tables
1133866416 M * robig sounds nice :)
1133866428 M * ag- hmm, i'm not quite sure about iptables, ebtables
1133866433 M * robig no problems with already used ports anymore :)
1133866440 M * ag- exactly
1133866478 M * ag- the only issue is the overhead involved in ngnet, this is why the current networking will be kept
1133866557 M * robig the use of both at the same time would be great.
1133866578 M * robig not all vservers need to hav full speed :)
1133866619 M * ag- it will be possible, we evoked the possibility: http://linux-vserver.org/NGNET-Development
1133866769 M * robig i dont unterstand how to set rlimits files
1133866833 Q * flock Ping timeout: 480 seconds
1133866892 M * ag- robig: http://linux-vserver.org/Resource+Limits
1133867089 M * SiD3WiNDR ag-: afaik iptables will work :)
1133867120 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il
1133867143 M * robig but I cant change the max number of sockets,right?
1133867849 M * Larsioa have anybody seen connection drops in vserver? I have a client thats hosting Ventrilo (Voice server) he tells me that every 30 mins or soo they are getting disconnected. If i host this ventrilo server outside the vserver (on the main) there are no problems.
1133868202 M * SiD3WiNDR odd
1133868608 M * robig i'm using teamspeak in a vsever with no problems.
1133868616 M * Larsioa Hmm odd
1133869026 Q * Duckx Ping timeout: 480 seconds
1133869592 M * ag- SiD3WiNDR: context-aware netfilter?
1133869677 M * ag- Larsioa: never experienced connection drops, maybe you can run it under strace and log the output to see what happens
1133870035 J * Duckx ~duckx@195.75.27.158
1133870780 Q * Millox Quit: leaving
1133872633 Q * flock Ping timeout: 480 seconds
1133872958 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il
1133874370 M * robig what of the following init scripts arnt needed in vservers?
1133874371 M * robig README             dns-clean        inetd        nviboot         single
1133874371 M * robig apache2            exim4            keymap.sh    ppp             skeleton
1133874371 M * robig atd                exim4.dpkg-dist  klogd        pppd-dns        ssh
1133874371 M * robig bootclean.sh       glibc.sh         makedev      procps.sh       stop-bootlogd
1133874373 M * robig bootlogd           halt             modutils     rc              sudo
1133874373 M * robig bootmisc.sh        hostname.sh      mountall.sh  rcS             sysklogd
1133874375 M * robig checkfs.sh         hwclock.sh       mountnfs.sh  reboot          umountfs
1133874375 M * robig checkroot.sh       hwclockfirst.sh  mountvirtfs  rmnologin       umountnfs.sh
1133874377 M * robig console-screen.sh  ifupdown         mysql        screen-cleanup  umountroot
1133874377 M * robig cron               ifupdown-clean   networking   sendsigs        urandom
1133874383 M * robig sorry for flodding :)
1133874519 M * daniel_hozac none are needed except for the services you want to run.
1133874590 M * robig what ist console-screen?
1133874626 M * TheSeer wild guess: screen?
1133874635 M * TheSeer (man screen ;-P )
1133874775 M * robig this has an init script? :) interesting
1133874804 M * ag- hwclock* too
1133874833 M * robig this i can remove i think
1133877785 Q * Larsioa Ping timeout: 480 seconds
1133879216 N * Bertl_zZ Bertl
1133879223 M * Bertl morning folks!
1133879326 M * robig hi Bertl *laola*
1133879387 M * Bertl robig: how are things going?
1133879415 M * robig little problems here and there ;)
1133879609 M * robig i havent tested the new kernel now... :) you can wish me good luck for tomorrow :)
1133879676 M * robig one question: is there a solution for restricting diskspace to a user of a vserver?
1133879699 M * robig for exampple over a virtual disk in a file?
1133880381 Q * shedi Quit: Leaving
1133880383 M * robig no idea?
1133880447 M * Bertl what do you mean by 'user'?
1133880484 M * Bertl linux-vserver supports so called 'disk limits' on shared partitions
1133880506 M * Bertl they are an upper limit to how much disk space an entire guest can take up
1133880541 M * robig no, i mean an system user of the guest os
1133880594 M * Bertl what is a system user?
1133880607 M * robig i think quota can do such a disk limit per user, buts not working with vserver kernel patch, right?
1133880628 M * robig adduser in guest os creates an system user :)
1133880629 M * Bertl quota is working quite fine, you just have to set it up properly
1133880658 M * Bertl for 2.6 kernels you need to have a separate partition for each quota guest
1133880675 M * Bertl the 2.4 tree also supports shared quota (quota on a shared partition)
1133880705 M * robig i simply want that user xyz in vserver vs01 can only save files about 50mb
1133880714 M * Bertl http://linux-vserver.org/Quota+and+Disk+Limits
1133880718 M * Bertl http://linux-vserver.org/Standard+non-shared+quota
1133880736 M * robig thats the problem.. my vservers are running on an softraid
1133880750 M * Bertl so?
1133880763 M * Bertl software raid works pretty fine with quota
1133880764 J * Ben81 ~Ben81@tipi0e.lri.fr
1133880770 M * Bertl welcome Ben81!
1133880809 M * robig do i have to separate the users into an partition one by one?
1133880819 M * Ben81 Hi Bertl 
1133880835 M * Ben81 I have some question about hard_sched :)
1133880839 M * Bertl robig: not the users, but the guests
1133880875 M * Bertl robig: or alternatively, if you control all the guests (root) then you can simply use different uids across the guests
1133880894 M * Ben81 why when i launch a big_process in a VM with the hard_sched set, the vtop show that only 25% of the cpu is used ?
1133880912 M * Ben81 (used by this process)
1133880918 M * Bertl because the default setup (for the scheduler) is roughly 25%
1133880929 M * Ben81 hoo ok
1133880937 M * Ben81 can i change this ? 
1133880952 M * Ben81 maybe u have an url  :)
1133880952 M * Bertl sure, check out: http://linux-vserver.org/Scheduler+Parameters
1133880965 M * Ben81 i'll see that thx
1133880972 M * Bertl you're welcome!
1133880996 M * Bertl okay, translocating now ... will be back in 40mins ...
1133881004 N * Bertl Bertl_oO
1133881185 M * robig bye
1133881191 M * robig cu l8r
1133881212 Q * robig Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org
1133881510 M * harry my first mail sent to vserver list!!!!!! :)
1133881521 N * nox_ nox
1133881689 M * nox what is the right /dev/pts entry in fstab (host and client) ?
1133881703 M * nox can't find any doc about pts security
1133881741 M * nox i get "Cannot open your terminal '/dev/pts/6' - please check."
1133882039 Q * flock Ping timeout: 480 seconds
1133882246 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il
1133883064 N * Bertl_oO Bertl
1133883068 M * Bertl back now ...
1133883072 M * nox wb Bertl 
1133883091 M * Bertl hey nox! LTNS!
1133883101 M * Bertl harry: congrats!
1133883121 M * Bertl nox: check group and user permissions
1133883150 M * nox Bertl: yes and back with newbish question ... shame on me
1133883161 Q * Duckx Ping timeout: 480 seconds
1133883206 M * nox thx Bertl.. permissions on the host ?
1133883291 M * nox it is 640 
1133883356 M * Bertl so that means that your user has to be in the proper group (probably tty)
1133883368 M * Bertl and no, the pts on the guest :)
1133883388 M * nox there are no entries in the guest
1133883500 M * nox is it save to create them ?
1133883512 M * nox hmm error still the same then
1133883558 J * Johnnie ~john@acs-24-154-53-16.zoominternet.net
1133883615 M * nox they are not shown from inside the guest
1133884245 Q * flock Ping timeout: 480 seconds
1133884406 M * Bertl nox: when the guest is started, the tools will mount the pts filesystem on /dev/pts
1133884421 M * Bertl what you should see by default is /dev/ptmx and /dev/pts dir
1133884445 M * Bertl nox: if you then 'allocate' a terminal, you will get a new entry in /dev/pts
1133884466 M * Bertl nox: what and how do you test/work/try?
1133884479 M * nox  none on /dev/pts type devpts (rw) is what mount shows me in the guest
1133884489 M * Bertl okay, that's fine
1133884517 M * nox it tried to open a screen session and to login over ssh
1133884535 M * Bertl open a screen session where?
1133884551 M * nox after entering from the host
1133884558 M * Bertl well, that's the fault
1133884558 M * nox in the guest
1133884575 M * Bertl if you use the 'enter' you are half outside, and half inside the guest
1133884597 M * Bertl i.e. the terminal will belong to the host, and you won't get access to it (for security reasons)
1133884610 M * nox stdin: is not a tty   is what ssh say even with -tt
1133884620 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il
1133884663 M * nox Bertl: ah ic (enter) and now it seems logical
1133884670 M * Bertl nox: what distro/tools/kernel?
1133884722 M * micah the OpenVZ concept of UBC (User BeanCounters), sounds not too different from the token bucket limits that Linux-vserver has
1133884739 M * nox 2.6.13.2-vs2.0.1-pre2 / debian testing /0.30.208-4
1133884749 M * Bertl micah: yes, it is basically identical, although they use it for more resources
1133884770 M * Bertl micah: it's also slightly more intrusive
1133884785 M * micah nox: the kernel-patch and util-vserver from unstable is more up-to-date
1133884828 M * Bertl nox: as who do you try to ssh into the guest? an user or root?
1133884882 M * nox Bertl: root
1133884901 M * Bertl and you try to reach guest root, right?
1133884914 M * nox yes i did a portforward
1133884935 M * Bertl port forward?
1133884955 M * Bertl anyway, check the config in your guest (the sshd one)
1133884964 M * nox yes the guest are nated since I only have one fixip
1133884967 M * Bertl and make sure to disable the UseLogin thingy
1133885006 M * nox ok do this later iam off for about 2hours
1133885021 M * nox Bertl: thx alot again 4 ya great support
1133885041 M * Bertl k, np
1133885044 M * Bertl you're welcome!
1133885105 J * shedi ~siggi@inferno.lhi.is
1133885502 M * Bertl welcome shedi!
1133885516 M * shedi thank you Bertl
1133885614 Q * eyck Quit: leaving
1133886150 Q * flock Ping timeout: 480 seconds
1133886246 J * eyck eyck@81.219.64.71
1133886290 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il
1133886445 Q * Ben81 Quit: Leaving
1133887398 J * stefani ~stefani@superquan.apl.washington.edu
1133887844 Q * flock Ping timeout: 480 seconds
1133888019 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il
1133888292 M * derjohn Is there a special trick to give a guest a RFC1918 IP? I mean if the host has a public IP and I want guests to NAT?
1133888345 M * SiD3WiNDR that works fine
1133888350 M * SiD3WiNDR but you need to enable snat on those
1133888357 M * SiD3WiNDR or something.
1133888361 M * SiD3WiNDR it's documented somewhere
1133888434 M * Bertl derjohn: just use -j SNAT --to <public-ip>
1133888501 M * derjohn will the 192er IP show up in other hosts arp cache (same collision domain) ?
1133888515 M * michal_ Bertl: you have that unsuported airport in your powerbook ? 
1133888518 M * michal_ hey btw :)
1133888544 M * SiD3WiNDR derjohn: only on the host itself afaik
1133888555 M * Bertl derjohn: nope, the snat will change the private ip to the public one and track the connection (i.e. reconvert on the other direction)
1133888584 M * Bertl michal_: guess no, my airport works quite fine >(
1133888607 M * derjohn thx, I wasnt sure if that works if the IP is in the same machine. I added that factoid to the wiki ;)
1133888642 M * Bertl k, btw, it seems that 'recent' masquerading also will handle the snat
1133888649 M * michal_ Bertl: lucky man, anyway, i've just got my airport working good with new driver (broadcom so it wasn't possible for now)
1133888654 M * Bertl derjohn: (this is unverified though)
1133888666 M * michal_ but reverse-eng driver is there now :)
1133888680 M * Bertl michal_: well, mine is orinoco based, so it worked out of the box
1133888691 M * SiD3WiNDR mine is full of planes
1133888704 M * Bertl lol
1133888716 M * ag- michal_: airport extreme has been reverse-engineered completely yesterday :)
1133888764 M * michal_ nice
1133888766 M * ag- it's ready and functional now, yaha!
1133888774 M * michal_ i am on that channel now, yeah
1133888796 M * michal_ so far i had to use cisco airport pcmcia under linux
1133888806 M * michal_ that's _the_ news anyway :]
1133888824 M * Bertl okay, off for a while, back later ...
1133888829 N * Bertl Bertl_oO
1133890033 J * Jamal ~Jamal@p548CC4F5.dip0.t-ipconnect.de
1133890534 Q * flock Ping timeout: 480 seconds
1133890718 Q * mountie Quit: LUNCK!
1133890986 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il
1133892056 J * mountie ~mountie@CPEdeaddeaddead-CM000a739acaa4.cpe.net.cable.rogers.com
1133892248 J * infowolfe ~infowolfe@209-112-219-176-cdsl-rb1.nwc.acsalaska.net
1133892329 J * gerrit ~gerrit@outside0.osdl.org
1133892391 N * Bertl_oO Bertl
1133892396 M * Bertl back now
1133892409 M * Bertl welcome Jamal! gerrit!
1133892423 M * Bertl wb flock! mountie!
1133892876 J * psg ~psg@zeb-att-nat.cooldog.com
1133892896 M * Bertl welcome psg!
1133892902 M * psg Herbert --- how are you?
1133892911 M * Bertl fine thanks! and zou?
1133892915 M * Bertl *you
1133892951 M * psg A bit harried --- I thought I had this "in the can" when I went to bed last night,now it turns out there's a mysterious problem ...
1133892954 M * Bertl funny issue the sybase thingy
1133893001 M * psg I just tried the 2.6.14.3-vs2.1.0-rc9, and found that it requires the Legacy API and Dynamic Contexts in the .config, so I'm rebuilding again.
1133893030 M * Bertl yup, you might want to stick to the stable branch for now
1133893038 M * Bertl (unless you need devel features)
1133893106 M * psg Well, I'll give this one a try first.  It seems to bomb out quite reliably if there is a problem, so if it makes it through the load db one time, I'll feel fairly confident.
1133893129 M * Bertl k, btw, do you know what it is trying to tell with that error?
1133893151 M * Bertl compress::/wrk1/db1_dumps/db/lpdb5::04 looks like it is going through some userspace packer?
1133893164 M * psg Which one, the latest with the new build?  Vserver won't even start --- chbind complains....
1133893177 M * Bertl no, that's kind of expected :)
1133893227 M * psg Ah, sybase ... What's going on is that there are about 6 processes that are reading and decompressing 5 gzipped database dump files, and cramming the contents back into the DB.
1133893270 M * psg The DB log file also had one weird error, that I have never seen before, somrthing like "Engine 1 terminated by scheduler"
1133893323 M * psg If I did have to go back to an older revision, which one would you recommend?
1133893328 M * Bertl hmm .. could OOM be an issue?
1133893367 M * Bertl i.e. might it be possible that your guest hits some limits and/or the memory of the server is exhausted?
1133893375 M * psg I don't believe so.  The box has 8GB on it, and this process has been running every morning at 5AM for months.
1133893436 M * psg And it bombs out at unpredictable places in the process of reloading.  Not the same spot every time.
1133893437 M * Bertl okay, just asking dumb questions :)
1133893467 M * psg That's OK.  Believe me, I appreciate how helpful (and available) you always are.
1133893491 M * psg Let me see if the build is done yet ...
1133893510 M * Bertl the no-such-process looks strange/suspicious to me
1133893535 M * psg done ... give me a minute or two to load it and reboot.
1133893546 M * Bertl yeah, sure .. no need to hurry ...
1133893567 M * Bertl btw, it's my pleasure to help ...
1133893610 Q * dos000 Ping timeout: 480 seconds
1133894263 M * gerrit Hi Bertl!
1133894411 M * psg Bertl --- no joy; with 2.6.14.3-vs2.1.0-rc9 I get the same problem while reloading the DB.
1133894431 M * Bertl okay, so it's kind of reproducible ...
1133894434 M * psg If you are up for it, I can rebuild with vserver debugging enabled
1133894438 M * Bertl (which isn't that bad)
1133894453 M * Bertl well, the questions we have to ask here are:
1133894456 M * psg It bombed out fairly quickly this time.
1133894466 M * Bertl - what would happen with 2.6.14.3 vanilla?
1133894490 M * Bertl - what actually happens?
1133894505 M * Bertl - and can we narrow it down?
1133894512 M * psg Whew ... good question.  This could be VERY hard to set up a test, since everything is placed where the vserver needs it.
1133894527 J * dos000 ~dos000@CPE00119572fd49-CM00137186e53a.cpe.net.cable.rogers.com
1133894528 M * Bertl let me make a suggestion how to proceed>
1133894536 M * psg I might be able to set it up with symlinks...
1133894537 M * psg ok
1133894546 M * Bertl I assume that your system is not in production, right?
1133894591 M * psg Hmmmm ... it sort of is ... in use by developers, and salepeople to get reports.  We can beat on it for a few hours, but not much more.
1133894604 M * Bertl okay, so I would first do the following:
1133894638 M * Bertl - compile a vanilla 2.6.14.3, and a 2.0.1-rc* kernel with your current .config file (used via make oldconfig)
1133894667 M * Bertl - have a test run with 2.0.1-rc* first, just to verify that the issue is still there
1133894691 M * Bertl - if it is there, then boot with the vanilla kernel and simply do a chroot into the guest
1133894725 M * Bertl just to avoid another compile run in all cases, enable the debugging stuff I will list now
1133894729 M * psg Ah ... of course .... ok, rc* ?
1133894731 M * psg ok
1133894767 M * Bertl http://vserver.13thfloor.at/Experimental/patch-2.6.14.3-vs2.0.1-rc5.diff
1133894780 M * Bertl CONFIG_DEBUG_KERNEL=y
1133894787 M * Bertl CONFIG_DEBUG_BUGVERBOSE=y
1133894787 M * Bertl CONFIG_DEBUG_INFO=y
1133894794 M * Bertl CONFIG_VSERVER_DEBUG=y
1133894820 M * Bertl CONFIG_KALLSYMS=y
1133894820 M * Bertl CONFIG_KALLSYMS_ALL=y
1133894851 M * Bertl best way is probably to add them by hand before you do the 'make oldconfig'
1133894861 M * Bertl (and verify that they are there afterwards)
1133895001 M * psg ok ... working on it now
1133895179 T * Bertl http://linux-vserver.org/ | latest stable 2.0, 2.0.1-rc5, 1.2.10, 1.2.11-rc1, devel 2.1.0-rc10 | util-vserver-0.30.209 | He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the wiki, and we'll forget about the minute ;) 
1133895210 M * Bertl -rc10 (new default for dynamic xid, experimental cfq stuff)
1133895451 M * psg Any objections to my doing 'make oldconfig' and then updating and checking with 'make gconfig'  ???
1133895504 M * Bertl nope
1133895510 M * psg thx
1133895538 M * Bertl sidenote: checking with grep is probably simpler :)
1133895575 J * sebi ~sebi@Fd6a1.f.strato-dslnet.de
1133895584 M * Bertl welcome sebi!
1133895613 M * psg Turn on "Vserver History Tracing" ?
1133895645 M * Bertl not required I guess
1133895671 Q * sebi_ Ping timeout: 480 seconds
1133896403 P * stefani I'm Parting (the water)
1133896451 M * psg building the vs2.0.1 package now.
1133897474 M * psg testing db load now ... keep your fingers crossed ...
1133897538 M * Bertl hehe
1133897906 M * psg It's made it halfway through --- about 10GB so far.  At least 50% farther than any previous runs.  I'm beginning to hope ...
1133897923 M * psg I can't watch it .... <LOL>
1133897947 M * psg I have to look now.
1133897985 M * psg 13GB now.
1133898015 J * independence ~independe@dsl-80-252-185-139-dynamic.areteadsl.se
1133898029 M * psg If this does work (knock on wood), I can do some testing for you after business hours.
1133898105 M * independence Hi, I'm having som problems with routing and internet access on a guest system, the thing is that the host system is also my internet router. Is this a problem anyone around here can help me with, or should I turn somewhere else?
1133898195 M * Bertl nope, that is a common issue/misunderstanding of linux routing/masquerading/nat
1133898195 M * Bertl independence: so we can help you here ...
1133898226 M * Bertl independence: do you use private addresses for the guest(s)?
1133898242 M * independence Um, my guest systems ip is 192.168.1.12
1133898252 M * independence And the router has the ip-adress 192.168.0.1
1133898263 M * independence (router = host system)
1133898282 M * Bertl okay, and where/how do you get public ip(s)?
1133898301 M * independence I get one public ip from my isp via dhcp
1133898306 M * Bertl some DSL router or so?
1133898321 M * Bertl or is that public IP on the host too?
1133898327 M * independence well, the host system is the router, but dsl yes
1133898350 M * Bertl the important question is, do you have a public ip assigned to your host?
1133898353 M * independence the host system has a network interface with the public ip
1133898357 M * independence yes
1133898362 M * Bertl okay, that clarifies ...
1133898402 M * Bertl you have to understand that the guests share the network stack with the host
1133898406 M * independence I'm very confused with everything right now, sorry.. :)
1133898415 M * independence Hmm, okay
1133898437 M * Bertl so, basically the guest tries to send out a packet, but uses the private ips
1133898451 M * Bertl of course, this will not be routed/handled properly
1133898467 M * independence Yes, that I understand
1133898477 M * Bertl the solution is simple, just add an SNAT rule like this:
1133898507 M * Bertl iptables -t nat -I POSTROUTING -s 192.168.1.12 -j SNAT --to <public-ip>
1133898510 M * independence I've even tried setting up a second default gateway with iproute2 that is in a table (I think it's called), and that table has a rule that says to only handle packages from 192.168.1.0/24
1133898518 M * independence Oh..
1133898532 M * Bertl forget the special routing, it's not required :)
1133898590 M * psg Bertl - the DB reload worked fin with vs2.0.1-rc5.  Email me later if you have any tests for me to run to help find the problem in 2.1.0-rc9  -- and thanks again!
1133898601 M * psg fin = fine
1133898607 M * Bertl psg: you're welcome! 
1133898629 M * Bertl psg: just let me know when you have time to test ...next step is to get a vs2.1.0 running with the same config
1133898665 M * independence Bertl: Yay! It works! Thanks a lot! :D
1133898689 M * independence Too bad I didn't have to use all those really cool iproute2 tools though.. ;)
1133898850 M * Bertl independence: lol
1133898872 M * Bertl independence: you're welcome! feel free to hang around ...
1133898923 Q * FireEgl Ping timeout: 480 seconds
1133898936 M * Jamal I have rented a vserver (i.e. only a guest system) from a web hosting company and would like to run ISPConfig (free alternative to Confixx, Plesk) on it. ISPConfig requires quota among other things. I am just a linux beginner so I thought before I try it and do something stupid I'd ask here if quota works inside a vserver? Thanks.
1133898970 M * Bertl Jamal: yes it works (with some limitations)
1133898981 M * daniel_hozac but only if configured properly on th ehost ;)
1133898983 M * daniel_hozac +spelling
1133899009 M * Bertl Jamal: per context quota on a shared partition is not implemented on the 2.6/vs2.0 version yet
1133899042 M * Bertl Jamal: quota inside a guest on a separate partition (e.g. lvm, evms, loop) works fine
1133899219 M * Bertl Jamal: does that answer thy question?
1133899421 M * Jamal The thing is I have no clue whatsoever about quota and I don't know what ISPConfig exactly needs quota to do so I probably have to ask someone from ISPConfig what functions they need and then get back to you to ask if that specific function is supported in a vserver. Thank you so far.
1133899674 J * FireEgl Atlantica@Atlantica.US
1133899984 M * Bertl Jamal: should work, and if not, you can always adapt ISPConfig (if zou got the source)
1133899994 M * Bertl wb FireEgl!
1133900321 M * Jamal Bertl: OK, I'll see what the ISPConfig folks say about it and then might just give it a shot.
1133900435 M * Bertl great, let us know how it goes ...
1133900455 Q * flock Remote host closed the connection
1133900527 M * Jamal will do
1133900647 J * robig ~robig_m@80.243.33.107
1133900651 M * robig re
1133900655 M * Bertl wb robig!
1133900735 M * robig hi ive got a problem with reached max open sockets. how can I change this?
1133900754 M * Bertl increase the number of available sockets?
1133900783 M * robig yes
1133900816 M * Bertl what's the message you get?
1133900852 M * robig i get no message
1133900870 M * Bertl hmm, so how do you know you reached a limit then?
1133900887 M * robig # cat /proc/virtual/49153/limit
1133900887 M * robig SOCK:          132             132              -1           0
1133900909 M * Bertl shows me that you still use dynamic ids :)
1133900923 M * Bertl and that your context has 132 sockets open right now
1133900937 M * robig yes because i dont know how to change that
1133900947 M * Bertl what? the dynamic ids?
1133900951 M * robig yes
1133900967 M * robig static ids are better?
1133900976 M * Bertl just add a file named 'context' to your config tree, containing a single (unique) number
1133901000 M * Bertl (or specify iy at guest creation time with --context <id>
1133901018 M * Bertl (numbers are between 2 and 49151(
1133901111 M * robig ok thanks
1133901113 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il
1133901122 M * robig now to the sokets problem
1133901129 M * Bertl yes, what is the problem?
1133901148 M * robig can i increase the available sockets of one vserver?
1133901185 M * robig were getting problems when many users are connected
1133901211 M * Bertl your config does not limit them in any way (-1 as limit)
1133901231 M * Bertl so I wonder why you conclude that it _is_ limited?
1133901252 M * robig oh
1133901254 M * Bertl (just trying to figure where you see limits)
1133901256 M * robig okay :)
1133901295 M * robig is 132 open sockets a high value?
1133901318 M * Bertl depends, I'd say it should be fine for a server (apache or tomcat)
1133901361 M * Bertl what problems do you get when many users are connected?
1133901385 M * robig the site is quiet slow
1133901437 M * Bertl do you impose cpu limits to the guest?
1133901480 Q * lilo Remote host closed the connection
1133901485 M * robig no i dont think so :)
1133901498 J * lilo ~lilo@lilo.usercloak.oftc.net
1133901546 M * robig the problem is.. our server admin has left the team, now i must do the server/vserver stuff :)
1133901561 M * Bertl robig: ah, i.c.
1133901577 M * Bertl robig: well, could you run the testme.sh for me and upload the output somewhere
1133901589 M * Bertl (e.g. pastebin.com) I might be able to provide some hints
1133901656 M * robig where do i find that file?
1133901722 M * Bertl sec
1133901744 M * Bertl http://vserver.13thfloor.at/Stuff/SCRIPT/testme.sh
1133901843 M * robig Bertl qry
1133901933 M * Bertl yup, so I'd suggest to update to 2.6.14.3 with 2.0.1-rc5
1133901950 M * robig ive done so
1133901954 M * Bertl and give that a try with your current load, maybe fine tuning a few kernel settings would not hurt
1133901975 M * robig if you can help me with that :D
1133901995 M * robig the new kernel is compiled, and waits for a test :)
1133902016 M * robig but i cont reboot with 25 people online :)
1133902024 M * robig cant
1133902034 M * Bertl okay, let me know how the 'new' kernel handles the load ...
1133902043 M * Bertl we can then have a look at optimizations
1133902054 M * robig ok thanks
1133902065 M * Bertl you're welcome! have fun!
1133902090 M * aba Bertl: do you have some moment of time for me?
1133902185 M * Bertl sure, roughly 10 minutes
1133902202 M * aba good :)
1133902223 M * aba You have some good checker code like "does vserver work? can one escape from the chroot?" ...
1133902254 M * Bertl testme.sh, testfs.sh and friends?
1133902255 M * aba however - it's a bit hard for "normal" people to use it, because it's just at your http-server, and doesn't indicate which one is current, etc.
1133902258 M * aba yes
1133902289 M * aba so, I'm considering whether it would be a good idea to make a tar-ball with them as "vserver-checkers" or so
1133902306 M * Bertl feel free to start such a project
1133902325 M * aba well, I cannot do it unless you give me enough hints and help :)
1133902329 M * Bertl I already suggested many times that this would not only be useful but also important
1133902335 M * Hollow aba: you could integrate it in the new util-vserver
1133902358 M * Bertl so you can assume that I will help wherever possible
1133902376 M * aba Hollow: well, it should be independend of the util-package, because you should be able to run the scripts w/o upgrading the utils ...
1133902393 M * aba (and, I think it would be great to install such a package inside the guests, and run some scripts by cron)
1133902403 M * aba Bertl: ok, very good. I'll do that as my next fun project than
1133902424 M * Bertl aba: IMHO some tests will require certain setups, so I'd basically aim for a 'special' QEMU disk image, which can do automated testing
1133902459 M * Bertl i.e. will allow to run all tests and report any unusual stuff
1133902476 M * Bertl you can then plug in the tools to get them tested, as well as new kernels
1133902487 M * aba Bertl: ok, that's more like a regression test
1133902494 M * Bertl (kind of regression suite, yes)
1133902510 M * aba I was thinking more on "make sure everything works, even after security kernel upgrades, reboots and the like", i.e. regression for admins in the wild
1133902519 M * aba Probably both is useful and necessary, yes.
1133902529 M * Bertl maybe they can be combined ...
1133902560 M * Bertl usually the testme.sh is sufficient to show typical admin issues
1133902579 A * aba will try to learn more now ...
1133902592 M * Bertl they would not need fs tests and quota testing or so, if it was already done at kernel release time
1133902646 M * aba well, if they do their own kernel, they might
1133902697 M * Bertl yes, of course, but in this case, the regression test would be the best solution anyway ...
1133902712 M * Bertl okay, off for a meeting now .. back later
1133902720 N * Bertl Bertl_oO
1133902720 M * Hollow Bertl: what about the persistant nx patch?
1133902723 M * aba yes. Well, we should have a paper describing what is what ...
1133902724 M * Hollow ic..
1133902726 M * Hollow cu ;)
1133907451 Q * robig Ping timeout: 480 seconds
1133907873 M * [MUPPETS]Gonzo which version of util-vserver is recommended at the moment?
1133907910 M * [MUPPETS]Gonzo latest? 0.30.209?
1133907930 M * mnemoc yes
1133907933 M * [MUPPETS]Gonzo thanks
1133908130 Q * psg Quit: 
1133908353 Q * aba Ping timeout: 480 seconds
1133908689 N * Bertl_oO Bertl
1133908710 M * Bertl Hollow: url?
1133909055 M * Bertl okay, translocating now ... back in 30 or so ...
1133909146 N * Bertl Bertl_oO
1133910212 J * Aiken ~james@tooax6-082.dialup.optusnet.com.au
1133913233 Q * entroposcope jupiter.oftc.net iridium.oftc.net
1133913233 Q * Pazzo jupiter.oftc.net iridium.oftc.net
1133913233 Q * FireEgl jupiter.oftc.net iridium.oftc.net
1133913253 Q * Jamal Quit:  HydraIRC -> http://www.hydrairc.com <- 100,000+ downloads can't be wrong
1133913384 J * FireEgl Atlantica@Atlantica.US
1133913384 J * entroposcope ~entroposc@user-0c992og.cable.mindspring.com
1133913384 J * Pazzo ~Pazzo@host130-250.pool8172.interbusiness.it
1133913454 Q * Pazzo iridium.oftc.net strange.oftc.net
1133913454 Q * entroposcope iridium.oftc.net strange.oftc.net
1133913503 Q * FireEgl iridium.oftc.net arion.oftc.net
1133913526 J * Pazzo ~Pazzo@host130-250.pool8172.interbusiness.it
1133913526 J * entroposcope ~entroposc@user-0c992og.cable.mindspring.com
1133913526 T * kinetic.oftc.net http://linux-vserver.org/ | latest stable 2.0, 2.0.1-rc5, 1.2.10, 1.2.11-rc1, devel 2.1.0-rc10 | util-vserver-0.30.209 | He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the wiki, and we'll forget about the minute ;) 
1133913591 J * FireEgl Atlantica@Atlantica.US