1182902983 J * rgl ~Rui@84.90.10.107
1182903098 M * onox Guy-: why do you want to create child processes in another context?
1182903214 Q * Radiance Remote host closed the connection
1182903435 M * Guy- onox: vserver exec?
1182903461 M * onox parent in host creates childs in guest?
1182903469 M * Guy- yes, that's what vserver exec does
1182903475 M * onox or parent in guest creates children in guest?
1182903480 M * onox (other guest)
1182903492 M * Guy- is that possible?
1182903511 M * onox don't want that to be possible :p
1182903557 M * Guy- neither do I
1182903562 M * onox good :)
1182903573 M * Guy- but I don't see what's wrong with being able to send signals to a child you were able to create :)
1182903584 M * onox i'm going to execute find / -iname "bed"
1182903610 N * onox onox`zZzZ
1182905106 J * DoberMann_ ~james@AToulouse-156-1-133-243.w90-30.abo.wanadoo.fr
1182905216 Q * DoberMann[ZZZzzz] Ping timeout: 480 seconds
1182907046 M * Guy- good night
1182909023 Q * bzed Quit: Leaving
1182913823 Q * Piet Quit: Piet
1182914403 N * Bertl_zZ Bertl
1182914412 M * Bertl Guy-: still around?
1182914419 M * Bertl morning folks!
1182915046 J * slack101 fgfdgfdg@cpe-71-74-77-84.insight.res.rr.com
1182915057 M * slack101 what up
1182915209 M * Bertl what down?
1182915296 M * slack101 lol
1182915302 M * slack101 Bertl, your not in bed yet ?
1182915317 M * Bertl unfortunately, I'm already up again :/
1182915321 M * slack101 lol
1182915332 M * slack101 im working hard to setup my VPS stuff real good 
1182915341 M * Bertl sounds good!
1182915363 M * slack101 i got the FTP recovery access working all good
1182915377 M * Bertl excellent
1182915388 M * slack101 it was actually a pretty good idea 
1182915743 M * slack101 Bertl, how many languages you speak ?
1182915776 M * Bertl only two, but I can understand a few more ...
1182915797 M * Bertl (not counting any computer languages, that is :)
1182916203 M * slack101 SPanish / Portuguese ? Bertl 
1182916272 M * Bertl I sometimes can partially understand spanish, yes
1182916281 M * slack101 Bertl, also you use Windows or Linux as desktop ?
1182916305 M * Bertl why would I use Windows? of course I use Linux
1182916646 M * slack101 Bertl, lately i have put some faith back into windows ;)
1182916666 M * slack101 WinSCP
1182916689 M * slack101 Putty are some pretty good apps ;)
1182916713 M * slack101 i had to use windows for a little while to find the best tools n stuff ot put together for my customers as most of them are going to be windows users 
1182917138 M * Bertl well, the choice is yours, I prefer a deterministic OS which doesn't spy on me :)
1182918263 M * slack101 Bertl, who is spying ?
1182918264 M * slack101 ;)
1182918283 M * slack101 no im jus using it right now so i can make it best for my customers ;)
1182920427 J * meandtheshell ~markus@85.127.102.3
1182920731 N * DoberMann_ DoberMann[PullA]
1182921598 M * slack101 Bertl, but yea i love linux but right now my windows distro beats my linux one ......linux one == KDE + etc etc 
1182921617 M * slack101 my linux would be like flux box etc 
1182921745 M * neuralis what's a "windows distro"?
1182921803 Q * rgl Ping timeout: 480 seconds
1182923089 M * slack101 neuralis, lol
1182923583 M * neuralis it was a serious question
1182923767 M * Bertl slack101: in what way? resource consumption? unintuitiveness? lack of control? :)
1182924168 M * slack101 Bertl, my windows is only using 80 mb right now im using an optimized edition 
1182924181 M * slack101 but yes i would rather be using fluxbox 
1182924747 M * Bertl slack101: on your handheld?
1182924763 M * slack101 lol ?
1182924768 J * DavidS ~david@p54811BDF.dip0.t-ipconnect.de
1182924771 M * slack101 Bertl,  which comment ?
1182924775 M * slack101 desktop
1182924791 M * Bertl well, I've never seen a recent windows below 2GB
1182924813 M * Bertl (not including stuff like office)
1182924906 M * slack101 i am talking ram useage 
1182924990 M * Bertl oh, hmm, well, my frontend machine has 64MB, so I think the usage is below that :)
1182925754 M * slack101 Bertl, yea im gonna set me up a linux with fluxbox it should be very low 
1182925759 M * slack101 35-40 mb proll y
1182926181 Q * sharkjaw Quit: Leaving
1182926196 J * sharkjaw ~gab@158.36.45.236
1182926342 M * eyck- hmm, linux + XFree86 + fluxbox fits in 24M ram machine quite comfortably
1182926347 M * eyck- don't know about xorg
1182926697 J * HeinMueck ~Miranda@dslb-088-064-028-081.pools.arcor-ip.net
1182927010 M * Bertl I guess I will take a nap now ... back later ..
1182927015 N * Bertl Bertl_zZ
1182930560 J * dna ~naucki@160-198-dsl.kielnet.net
1182931098 J * chand ~chand@212.99.51.254
1182931598 J * pflanze ~chris@77-56-91-62.dclient.hispeed.ch
1182931608 M * pflanze Hello.
1182931611 Q * ensc Ping timeout: 480 seconds
1182931664 M * pflanze Bertl_zZ: I've found out (at least part of) the reason Gnome applications are taking so long (more than 1 minute) to start up after reunification
1182931720 M * pflanze They scan the fonts directory and stat and open every file therein.
1182931744 M * pflanze And at least this vserver guest has:
1182931747 M * pflanze find -type f /usr/X11R6/lib/X11/fonts/|wc -l
1182931747 M * pflanze 3798
1182931756 M * pflanze files. So..
1182934804 J * pmenier ~pmenier@LNeuilly-152-22-72-5.w193-251.abo.wanadoo.fr
1182934873 Q * chand Quit: chand
1182935192 Q * neuralis Read error: Connection reset by peer
1182935294 J * neuralis ~krstic@solarsail.hcs.harvard.edu
1182935575 Q * Guy- charon.oftc.net unununium.oftc.net
1182935575 Q * pyquila charon.oftc.net unununium.oftc.net
1182935575 Q * kaner charon.oftc.net unununium.oftc.net
1182935575 Q * jannic charon.oftc.net unununium.oftc.net
1182935575 Q * transacid charon.oftc.net unununium.oftc.net
1182935575 Q * yang charon.oftc.net unununium.oftc.net
1182935575 Q * fosco charon.oftc.net unununium.oftc.net
1182935575 Q * vasko charon.oftc.net unununium.oftc.net
1182935575 Q * Loki|muh charon.oftc.net unununium.oftc.net
1182935575 Q * SadMan charon.oftc.net unununium.oftc.net
1182935594 J * Guy- ~korn@152.66.83.144
1182935594 J * pyquila gerbens@82.94.222.35
1182935594 J * kaner kaner@strace.org
1182935594 J * jannic ~jan@gasko.hitnet.RWTH-Aachen.DE
1182935594 J * transacid ~transacid@transacid.de
1182935594 J * yang ~yang@yang.sponsor.oftc.net
1182935594 J * fosco fosco@konoha.devnullteam.org
1182935594 J * vasko ~vasko@unreal.rainside.sk
1182935594 J * Loki|muh loki@satanix.de
1182935594 J * SadMan sadman@sadman.net
1182935662 Q * slack101 Quit: Leaving
1182935731 M * Guy- Bertl_zZ: I am now
1182937165 Q * SadMan galapagos.oftc.net unununium.oftc.net
1182937165 Q * Loki|muh galapagos.oftc.net unununium.oftc.net
1182937165 Q * vasko galapagos.oftc.net unununium.oftc.net
1182937165 Q * jannic galapagos.oftc.net unununium.oftc.net
1182937165 Q * pyquila galapagos.oftc.net unununium.oftc.net
1182937165 Q * Guy- galapagos.oftc.net unununium.oftc.net
1182937165 Q * transacid galapagos.oftc.net unununium.oftc.net
1182937165 Q * fosco galapagos.oftc.net unununium.oftc.net
1182937165 Q * kaner galapagos.oftc.net unununium.oftc.net
1182937165 Q * yang galapagos.oftc.net unununium.oftc.net
1182937187 J * Guy- ~korn@152.66.83.144
1182937187 J * pyquila gerbens@82.94.222.35
1182937187 J * kaner kaner@strace.org
1182937187 J * jannic ~jan@gasko.hitnet.RWTH-Aachen.DE
1182937187 J * transacid ~transacid@transacid.de
1182937187 J * yang ~yang@yang.sponsor.oftc.net
1182937187 J * fosco fosco@konoha.devnullteam.org
1182937187 J * vasko ~vasko@unreal.rainside.sk
1182937187 J * Loki|muh loki@satanix.de
1182937187 J * SadMan sadman@sadman.net
1182937212 Q * coderanger Ping timeout: 480 seconds
1182937847 N * pmenier pmenier_off
1182938203 N * Bertl_zZ Bertl_oO
1182938210 M * Bertl_oO back later ...
1182938878 M * derjohn Bertl_oO, gone? How many delta do  have to apply in order to make ioacct cleanly apply ?
1182939169 J * chand ~chand@212.99.51.254
1182939220 Q * chand 
1182939591 J * Piet hiddenserv@tor.noreply.org
1182939766 J * rgl ~Rui@84.90.10.107
1182940423 J * Piet_ hiddenserv@tor.noreply.org
1182940751 Q * Piet Ping timeout: 480 seconds
1182940805 J * ensc ~irc-ensc@p54B4ECF1.dip.t-dialin.net
1182940902 N * Piet_ Piet
1182944784 J * eSa| ~kvirc@ip-87-238-2-45.adsl.cheapnet.it
1182944914 Q * Aiken Quit: Leaving
1182945885 Q * eSa| Read error: Connection reset by peer
1182945895 J * eSa| ~kvirc@ip-87-238-2-45.adsl.cheapnet.it
1182946009 M * rgl udev cannot run inside a guest, correct?
1182946931 M * arachnist  first you'd need to allow a guest to create dev nodes
1182946962 M * arachnist but then you'll get another error if you have udev on your host
1182946975 M * arachnist (address already in use or something like that)
1182947711 M * rgl I see.  I really don't have a need for running it inside the guest.  I'm just asking that because I want to remove the package :D
1182947993 N * pmenier_off pmenier
1182950082 M * daniel_hozac derjohn: why doesn't it apply cleanly? it looks like a rather inobtrusive patch to me.
1182950125 M * derjohn at least Bertl_oO told me to need sock-fix02. that changes there reject, but doesnt solve it.
1182950178 M * derjohn i didnt do into it yet, because I think i just need to patch in all deltas since some special date. 
1182950185 M * derjohn (after 2.2.0 release ?)
1182950356 Q * pflanze Ping timeout: 480 seconds
1182951150 J * vserveraddict ~vserverad@LPuteaux-151-41-11-129.w217-128.abo.wanadoo.fr
1182951606 M * vserveraddict hi there
1182951615 J * fatgoose ~samuel@76-10-153-123.dsl.teksavvy.com
1182951664 M * daniel_hozac hello
1182951672 M * vserveraddict i have a quick little question, do any one is able to create a vserver that is not the same than the host (debian doesn't count). Exemple debian host, centos guest ?
1182951680 M * vserveraddict i fail each time
1182951782 M * vserveraddict for exemple:
1182951782 M * vserveraddict vserver cent0Sclone \ 
1182951782 M * vserveraddict build -m yum -n cent0Sclone \ 
1182951782 M * vserveraddict --force --context 40352 \ 
1182951782 M * vserveraddict --hostname web26clone.aqueos.net \ 
1182951783 M * vserveraddict --interface 351=eth0:10.246.53.51/24 \ 
1182951783 M * vserveraddict -- -d centos4 -- 
1182951793 M * vserveraddict result in No dynamically linked rpm binary found; exiting...
1182951963 M * daniel_hozac and, do you have rpm and yum installed?
1182952138 M * vserveraddict humm lets say no (blush)
1182952143 M * vserveraddict so aptitude install yum
1182952163 M * vserveraddict now it seems to be doing things, obviously  it work better...
1182952192 M * vserveraddict You are using a version of yum which is insecure and broken in chroot 
1182952192 M * vserveraddict related operations; either apply the patches shipped in the 'contrib/' 
1182952192 M * vserveraddict directory of util-vserver, or ask the author of yum to apply them
1182952209 M * vserveraddict humm ... :)
1182952223 M * vserveraddict it fail with
1182952240 M * vserveraddict RTNETLINK answers: File exists
1182952351 Q * sharkjaw Remote host closed the connection
1182952377 P * SMMDI satan made me do it
1182952386 M * daniel_hozac that just means you told it to use an IP address already assigned to the host.
1182952395 J * ema ~ema@rtfm.galliera.it
1182952398 M * vserveraddict oooh
1182952398 M * vserveraddict ok
1182952403 M * vserveraddict yes this is true
1182952411 M * vserveraddict i want to share it
1182952473 M * daniel_hozac so you should've used nodev:, not 351=eth0:.
1182952482 M * daniel_hozac (assuming you have recent enough utils to support that syntax)
1182952503 M * vserveraddict ok so --interface nodev would do the trick
1182952508 M * vserveraddict thanks for the info
1182952538 M * daniel_hozac --interface nodev:<IP>/<prefix>, yes.
1182952569 M * vserveraddict ok
1182952587 M * daniel_hozac if your utils are too old, just drop the nodev:.
1182952600 M * daniel_hozac (or upgrade them)
1182952616 M * vserveraddict i am with the last one linked on the site :)
1182952740 J * JimBobDoe ~satan@89.155.106.6
1182952761 Q * gerrit Ping timeout: 480 seconds
1182952799 M * vserveraddict thanks i start a new build and test
1182953623 Q * rgl Ping timeout: 480 seconds
1182953943 J * cedric ~cedric@80.70.39.67
1182954996 M * vserveraddict ok it worked even with the vyum message
1182955013 M * vserveraddict thanks for pointing the obvious misses i had
1182955014 M * vserveraddict ;)
1182955014 M * Piet hey
1182955106 M * Piet does the lack of 'E' in the 'tag' column on the 'CPU' row in that table at http://linux-vserver.org/Resource_Limits#Linux_Resource_Limits mean that this limit cannot be enforced or that it is not enforced by default?
1182955173 M * Piet also, to make sure I understand this correctly: the limits ni this table are virtualized under linux-vserver and can be applied to contexts, not just single processes, right?
1182955192 M * Piet ni=in
1182955703 M * vserveraddict i think nothing mean E  as those limits are basic ones but i am not sure :)
1182955795 M * Piet hmm, so why wouldn't it say 'E' there?
1182955928 J * arekm arekm@carme.pld-linux.org
1182955945 J * coderanger ~coderange@c-65-96-210-168.hsd1.ma.comcast.net
1182955972 M * arekm hello, who is able to give access to "mbox" archives here http://archives.linux-vserver.org/ ? (or other archive)
1182956028 M * arekm and the other question, are there any (python) bindings for vserver stuff?
1182956139 M * daniel_hozac sladen should be the one.
1182956158 J * Radiance e2633164c8@halt.1984world.eu
1182956172 M * daniel_hozac Piet: the CPU limit is not enforced.
1182956186 M * daniel_hozac Piet: since it doesn't make much sense to say "this guest can run for 100 ms, then it should die".
1182956230 M * daniel_hozac arekm: oh, and for python bindings, i guess you'd want to poke coderanger :)
1182956259 M * daniel_hozac Piet: and yes, the limits listed as AE in that table are context wide.
1182956340 M * Piet daniel_hozac: oh, this indeed wouldn't make sense. i was hoping it would be more like "this guest can use X ms CPU time per minute" or so.
1182956364 J * kir_home ~kir@205.233.53.223
1182956373 M * daniel_hozac Piet: that's what the hard CPU scheduler is for. see http://linux-vserver.org/CPU_Scheduler
1182956426 M * Piet does it say anywhere on this page that the limits listed as AE in that table are context wide? Or can this be deducted from some other information given or referred to on this page?
1182956489 M * daniel_hozac well, it's a Linux-VServer page. the documentation only covers those aspects...
1182956491 M * Piet wow, the hard cpu scheduler forumlas look frightening
1182956585 M * Piet hmm, so the page only covers linux-vserver aspects, but not all rows in this table refer to contexts, but all of them refer to linux-vserver?
1182956597 M * daniel_hozac they all refer to contexts.
1182956600 Q * pmenier Read error: Connection reset by peer
1182956620 M * daniel_hozac there are just a bunch of them which are not implemented.
1182956631 M * Piet ah ok
1182956647 M * Piet thanks for your help
1182956708 J * pmenier ~pmenier@LNeuilly-152-22-72-5.w193-251.abo.wanadoo.fr
1182956793 M * Piet one more question: is there a way to limit I/O load on a per context basis, too?
1182956819 M * daniel_hozac not yet... the accounting is pretty much brand new.
1182956820 M * Piet I assume this is covered by the CPU scheduler, right?
1182956872 M * daniel_hozac well, not really.
1182956916 M * sladen arekm: there are mbox archives, and I'm happy to supply them with reason :)
1182956931 Q * kir_home Ping timeout: 480 seconds
1182956952 M * arekm sladen: mutt reading is what I love
1182956955 M * Piet does v2.0.2.1 already support CPU scheduling?
1182956984 M * daniel_hozac yes, but not fair sharing.
1182956991 A * arekm has question about rss limit. Does it count RAM only or swap, too?
1182957004 J * David1 ~david@p54812C8E.dip0.t-ipconnect.de
1182957037 M * sladen arekm: wow, the last time anyone asked was  200408.mbox.gz  apparently
1182957047 M * daniel_hozac RSS is resident set size, so it's RAM only.
1182957089 J * kir_home ~kir@205.233.53.223
1182957381 Q * DavidS Ping timeout: 480 seconds
1182957611 Q * kir_home Ping timeout: 480 seconds
1182957697 J * stefani ~stefani@flute.radonc.washington.edu
1182957813 N * pmenier pmenier_off
1182957842 Q * HeinMueck Quit: Aah!
1182957862 J * HeinMueck ~Miranda@dslb-088-064-028-081.pools.arcor-ip.net
1182958018 M * arekm sladen: well :-) Last time I was reading entire ML archive was ~2 years ago so.. 
1182958030 M * Piet daniel_hozac: are you aware of some scripts which apply the cpu scheduler principle on all vservers, possibly allowing one to equally distribute the available resources across all vservers, maybe with increased priorities on special vservers?
1182958042 M * Piet i.e., has any of this been scripted, yet?
1182958067 M * Piet it seems like this would be a very neccessary thing needed by more or less everyone running linux-vserver?
1182958157 M * daniel_hozac everyone has their own requirements for CPU limits..
1182958201 J * rgl ~Rui@84.90.10.107
1182958206 M * rgl hi
1182958289 M * daniel_hozac hello
1182958322 J * kir_home ~kir@205.233.53.223
1182958323 M * daniel_hozac Piet: note that you could also quite easily get that by only using the idle time scheduler.
1182958452 N * David1 DavidS|Vechta
1182958542 M * arekm what's the best way to do network traffic accouting for vserver? I see that there is something in /proc/virtual/XYZ/cacct but it would be probably bad way to do accounting
1182958591 M * arekm I'm thinking about just doing iptables based accounting on host
1182958658 M * Piet daniel_hozac: i was thinking of a configurable script, this would make it easily usable, and reusable. Is the idle time scheduler documented anywhere?
1182958703 M * daniel_hozac arekm: that works fine.
1182958712 M * daniel_hozac Piet: on that page...
1182958724 M * daniel_hozac Piet: you don't need a script though.
1182958725 M * arekm daniel_hozac: "that" means what? cacct or iptables?
1182958740 M * daniel_hozac arekm: well, both, but it was referring to iptables ;)
1182958772 M * rgl you guys known if its possible to get the microcode version of an intel cpu?
1182958780 M * Piet daniel_hozac: are you referring to "If the idle time ratio is the same for all contexts, the formula can be simplified:"?
1182958797 M * arekm daniel_hozac: but isn't cacct counting only outside--vserver_guest communication or everything that happens ofer INET for example? (so telneting from inside vserver guest to vserver guest itself is counted there)
1182958851 A * arekm needs vserver guest -- outside world accouting only
1182958970 M * Piet rgl: http://www.linuxfirmwarekit.org/
1182958994 M * daniel_hozac arekm: then iptables is what you want.
1182959000 M * Piet rgl: check the release notes for more info on the test
1182959024 M * daniel_hozac Piet: i'm referring to the sections on fair sharing and guarantees.
1182959076 M * Piet daniel_hozac: i get you now, thanks for your help
1182959308 M * arekm CTX is unsigned 32bit ?
1182959316 M * arekm .. int
1182959370 M * daniel_hozac it's signed. it's also limited to 2-65535.
1182959449 M * rgl Piet, thx.   From the docs I'm not yet seeing how can I get the microcode version, but I'll run it :D
1182959801 M * arekm Piet: debian has some script btw
1182959815 M * arekm Piet: http://ftp.debian.org/debian/pool/main/u/util-vserver/util-vserver_0.30.213-1.diff.gz util-vserver-0.30.213.orig/debian/examples/vschedcalc
1182960077 M * daniel_hozac heh, funny Debian has it. i think Hollow removed it from Gentoo :)
1182960129 M * Hollow hm.. we didn't port it to util-vserver, and we removed the gentoo-tools.patch ...
1182960158 M * arekm and debian doesn't have vlimit manual page which is weird thing in debian 8)
1182960171 Q * ensc Ping timeout: 480 seconds
1182960171 M * daniel_hozac there are no manual pages.
1182960176 M * daniel_hozac use <command> --help.
1182960235 M * arekm exactly and debian developers almost alwas written one if there was no man page upstream
1182960262 M * arekm hm, how to set sched_hard on running guest?
1182960275 M * daniel_hozac vattribute --set --xid <guest> --flag sched_hard
1182960355 M * arekm ok, [--flag [~!]<flag>] , what does ~ and ! do? ! is negation I guess but ~?
1182960362 J * ensc ~irc-ensc@p54B4F7B4.dip.t-dialin.net
1182960399 M * daniel_hozac same.
1182960413 M * daniel_hozac ~ is just the more logical one :)
1182960430 M * Hollow also ! is problematic in bash country iirc
1182960440 M * daniel_hozac yeah.
1182960458 J * bonbons ~bonbons@2001:5c0:85e2:0:20b:5dff:fec7:6b33
1182960474 M * daniel_hozac granted, as is ~ if you have users named sched_hard or another flag name :)
1182960484 M * Hollow yeah, bash sucks :)
1182960521 M * daniel_hozac i take it you're another one of those zsh fans? ;)
1182960650 M * Hollow no, i use bash :)
1182960661 M * daniel_hozac hehe, okay.
1182960725 Q * cedric Quit: cedric
1182960772 M * Piet ah great hint arekm, thanks
1182960902 Q * kir_home Ping timeout: 480 seconds
1182960976 A * arekm remembers reading about reboot being supported from guest starting from some newer vserver version. Does it actually reboot guest (stop && start) ?
1182961013 M * daniel_hozac "newer"? it's been supported in at least all of the 2.6 kernels.
1182961046 M * daniel_hozac even 1.2 has that.
1182961048 M * arekm fine, I was reading some old text then
1182961061 M * daniel_hozac s/old/ancient/
1182961078 M * arekm so.. does it stop and start guest at reboot? :)
1182961097 M * daniel_hozac well, yes. otherwise it wouldn't be a reboot, would it?
1182961128 M * arekm it could be simplified to shutdown only
1182961150 M * daniel_hozac that's what halt does.
1182961279 M * arekm daniel_hozac: http://pastebin.com/937626
1182961315 M * arekm does sysvinit need some modification?
1182961568 M * arekm hm, maybe init is not used here 8)
1182961701 M * arekm yeah, exactly, that was it
1182962115 M * Piet arekm: can you let me know how it's setup properly? i'm having issues with vservers not restarting automatically on debian, too
1182962193 Q * sladen Ping timeout: 480 seconds
1182962257 J * sladen paul@starsky.19inch.net
1182962261 M * arekm Piet: echo 'plain' > /etc/vservers/XYZ/apps/init/style
1182962272 M * arekm Piet: first stop; then do that; and then start again
1182962347 M * Piet thank you
1182962471 M * arekm is there something like vattribute --xid xyz show?
1182962529 M * arekm to get currently set attributes
1182962743 M * arekm hm, don't see such thing in /proc/virtual :/
1182963024 J * kir_home ~kir@205.233.53.223
1182963047 M * kir_home I'm sitting at the OLS, Linux-VServer presentation has just started :)
1182963057 M * kir_home anybody here who's also on the OLS?
1182963074 N * onox`zZzZ onox
1182963100 Q * dna Quit: Verlassend
1182963124 Q * yang Ping timeout: 480 seconds
1182963143 M * onox kir_home: OLS?
1182963148 M * arekm kir_home: ask a question for me ;-) 
1182963273 M * rgl is there a way to make the loopback interface appear inside a guest on the ifconfig listing?
1182963306 M * arekm rgl: the same way as other interfaces/ips I guess
1182963457 M * kir_home onox, Ottawa Linux Symposium. Herbert was going to talk but it looks like he had some problems with visa/tickets/etc.
1182963478 M * kir_home arekm, what's the question? :) 
1182963523 M * vserveraddict yep anyway multiple guest could not use lo  like a real loopback as it would be shared by all of them. If i remember you should create 127.0.0.2/3/4/5 etc.. no ?
1182963548 M * arekm kir_home: there is vattribute command that allows to set flags and capabilities for already running vserver guests. Now I'm looking for a way to get currently set flags/capabilities - is there such command or data in /proc?
1182963595 M * Hollow kir_home: ah, not the only one having visa problems? :)
1182963735 Q * pmenier_off Quit: pmenier_off
1182963736 M * arekm    vcmd -i 100 -BC ctx_create .flagword=^34^33^32^8 -- cpuhog :-) omg, black magic
1182963755 M * kir_home Hollow, oh yeah :) but it was not a problem for us to get Canada visa.
1182963766 M * Hollow hehe
1182963994 M * rgl vserveraddict, yup.  I assign them like: 127.0.0.<guest id>
1182964052 M * rgl arekm, you betcha!  what does that do? *G*
1182964066 M * arekm is cpu power scheduler on per context basis or on per process basis? for example I have 4 context running 1, 5, 100, 500 "cpuhog" processes respectively
1182964081 M * arekm now will each context get 25% of cpu?
1182964097 M * daniel_hozac kir_home: i think it was related to OLS paying for it.
1182964103 M * arekm rgl: running cpuhog with some cpu related settings
1182964128 M * rgl arekm, what is "cpu power schduler"?
1182964144 M * daniel_hozac arekm: use reboot -f.
1182964151 N * DoberMann[PullA] DoberMann
1182964156 M * arekm s/cpu power schduler/cpu scheduler/
1182964168 M * daniel_hozac arekm: and the flags/capabilities are in /proc/virtual/<xid>/status
1182964219 M * arekm daniel_hozac: ok, vattribute needs extending to parse that into human readable form 8-)
1182964229 M * daniel_hozac vserveraddict: no need to do that at all. 127.0.0.1 is rewritten to the guest's first IP address.
1182964288 M * daniel_hozac arekm: if you set up the scheduler for 25% each, then yes.
1182964320 M * arekm daniel_hozac: so setting 127.0.0.X as first ip would be nice because there would be no conflict in bind() for example (like bind(127.0.0.1) and then bind(first.ip))
1182964335 M * daniel_hozac hmm?
1182964370 M * arekm bind(127.0.0.1 port 50) and bind(12.12.12.12 port 50) will work at the same time?
1182964406 M * daniel_hozac Piet: to start guests on boot, do echo default > /etc/vservers/<guest>/apps/init/mark
1182964413 M * arekm if 127.0.0.1 is rewritten to first guest ip (12.12.12.12) then I would think that there will be conflict for bind()
1182964415 M * daniel_hozac arekm: yes.
1182964426 M * Piet herbert is another free software terrorist, those dont get visas lightly
1182964451 M * daniel_hozac but why would you bind to 127.0.0.1 and your public IP address separately?
1182964451 M * Piet thank you daniel
1182964471 M * daniel_hozac if you want to listen to all addresses, use INADDR_ANY.
1182964518 M * arekm daniel_hozac: there is for example rpcbind daemon which by default binds to 127.0.0.1 (and you can't turn off that without editing source code) + I want configure some public IP 
1182964539 M * daniel_hozac sounds broken to me.
1182964582 M * arekm daniel_hozac: there are some reasons for that I guess (rpcbind is newer replacement for portmap)
1182964635 M * arekm daniel_hozac: anyway back to more interesting cpu case. I'm interested what will happen by default in my cpuhog example? Will cpu scheduler give each 1) context of 2) process fair amount of cpu time?
1182964677 M * arekm s/of/or/
1182964716 M * daniel_hozac it depends on how you configure it.
1182964748 M * daniel_hozac but if you use the hard CPU scheduler, it will be context-based primarily.
1182964959 M * arekm I'm looking for easier way to get fair share per context behaviour. Is vattribute --set --xid <guest> --flag sched_hard enough for that?
1182964963 N * DoberMann DoberMann[PullA]
1182964977 Q * kir_home Ping timeout: 480 seconds
1182964996 M * arekm or do I need to play with vsched each time?
1182965041 M * daniel_hozac if the defaults are what you want, just setting the flag should be fine.
1182965139 M * arekm so to make me sure: the default with the sched_hard flag is fair share between context and without sched_hard the default is fair share between all processes in all contexts?
1182965196 M * daniel_hozac well, the defaults with sched_hard is that every context gets 25% of the CPU.
1182965215 M * arekm even if there is 10 contexts?
1182965225 M * daniel_hozac yes :)
1182965247 M * arekm uh. and 1 cpu or 25% of all available cpus?
1182965276 M * daniel_hozac 25% per CPU.
1182965318 Q * ema Quit: leaving
1182965375 M * arekm hm, but sched_hard flag is set per guest right? So few guests with that flag, few without it and... I'm lost then 8)
1182965400 M * daniel_hozac the guests with it will not be able to use more than 25% per CPU.
1182965443 Q * zLinux Remote host closed the connection
1182965683 M * arekm now let's assume that we have 2 guests, each guests has sched_hard set and that all bucket options set in a way that each guest can get max 60% of cpu (and we have only 1 cpu)
1182965731 M * arekm now first guest has 1 cpuhog process and second guest has 100 cpuhog processes. will these get 50% each ? :)
1182965750 M * arekm 50% of cpu time each
1182965778 M * daniel_hozac most likely the second one would get 60, and the first would get 40.
1182965835 M * arekm ups, that's bad
1182965898 A * arekm thinks about situation when there are 20 VPSes for 20 customers. Each customer gets for example max 50% cpu power limit. Now one bad customer can run thousands of processes and get all that 50% cpu for himself leaving 19 customers with the rest 50% for all 19
1182965940 M * arekm Limiting each customer to 100/20 (5) % doesn't look sane
1182965943 M * daniel_hozac so, uh, setup fair sharing instead?
1182965969 J * [BiG^BrotheR] ~Dr-DreaM@84.23.96.253
1182965994 M * arekm ehm, I see sched_hard and sched_prio only
1182966006 M * daniel_hozac yes, you need sched_hard for fair sharing.
1182966014 J * jakesol ~richardso@adsl-074-186-026-107.sip.mia.bellsouth.net
1182966019 M * daniel_hozac but you have to set the idle-time attribute in the scheduler.
1182966092 M * jakesol where can i obtain the util-vserver-0.30.213 rpm for CentOS 5 x86_64 ??
1182966105 M * arekm and then set limit to 5% (+ that idle time sharing) ? correct?
1182966210 M * daniel_hozac sure.
1182966239 M * daniel_hozac jakesol: i don't think there are any yet. you can build them yourself with rpmbuild -tb util-vserver-0.30.213.tar.bz2 though on your CentOS 5 x86_64 box though.
1182966262 M * daniel_hozac (i'm going to build some, just as soon as i get support for CentOS 5 into my buildsystem)
1182966481 M * arekm daniel_hozac: are some practical hints on how calculate/choose rate/tokens/interval written somewhere?
1182966500 M * daniel_hozac http://linux-vserver.org/CPU_Scheduler
1182966618 M * arekm I don't see practical hints there, like what values of bucket size are good/bad 
1182966645 M * daniel_hozac "good" and "bad" depend entirely on what you want to do.
1182966734 M * arekm like I have  2 x dual core cpu machine and X guests and I want to get fair sharing for all these guests 
1182966791 M * daniel_hozac lower values for rate/interval will give you a more interactive system, but with more scheduler overhead. higher values might make it seem sluggish, but it will be able to get more work done.
1182966855 J * James1 ~jltaylor@72.24.216.31
1182966864 M * arekm higher means 100, 10000, 10000000000 ?
1182966877 M * daniel_hozac depends on your HZ.
1182966916 M * daniel_hozac if your inteval is 2000, and your HZ is 100, that's 20 seconds between refills.
1182966927 M * daniel_hozac +r
1182967298 M * arekm what's fill-rate then? is this number of buckets put at each single refill ?
1182967338 M * arekm s/buckets/tokens/
1182967381 P * jakesol 
1182967477 M * daniel_hozac yes.
1182967485 M * daniel_hozac as that page explains.
1182967572 J * sauron ~zeus16384@12.20.4.100
1182967585 A * arekm adds vsched options in appropriate parts of that wiki then
1182967594 M * daniel_hozac no.
1182967598 M * daniel_hozac that belongs on another page.
1182967598 M * sauron hi all! could somebody tell me how I can check what vserver kernel patch version I have?
1182967604 M * daniel_hozac uname -a
1182967621 M * sauron daniel that doesn't...
1182967628 M * sauron i'd like the actual vserver patch ver
1182967651 M * daniel_hozac then you're not using a vanilla kernel with the vanilla patch intended for it.
1182967654 M * daniel_hozac so you're on your own.
1182967662 M * sauron i get 2.6.18-4-vserver-686
1182967675 M * daniel_hozac that means Debian, with 2.0.2.2-rc9.
1182967700 M * arekm http://oldwiki.linux-vserver.org/vsched+explained ok, this page is much nicer :)
1182967702 M * sauron mmm... I thought I can get stable kernel w/ 2.2 as a deb pkg
1182967713 M * daniel_hozac sure.
1182967721 M * daniel_hozac sid should have 2.6.21 packages.
1182967733 M * sauron really.... I check thanks a bunch.
1182967773 M * sauron be back.
1182967783 Q * sauron 
1182968038 M * Piet actually there's backported 2.6.21 packages
1182968060 M * Piet i.e. you can use these with debian etch
1182968066 M * Piet www.backports.org
1182968541 J * bzed ~bzed@dslb-084-059-114-192.pools.arcor-ip.net
1182969106 J * slack101 fgfdgfdg@cpe-71-74-77-84.insight.res.rr.com
1182969162 M * arekm ehm, to get fair sharing between 50 guests I should set interval = 2, fill_rate = 1. this doesn't look to sane, doing scheduling every 2 jiffie
1182969190 M * daniel_hozac that'd give each guest 50%.
1182969228 M * arekm so fill 1, int 50 :)
1182969258 M * daniel_hozac well, fill2 and int2.
1182969321 M * arekm fill2? aren't these 2 things for idle scheduling?
1182969345 M * daniel_hozac yes.
1182969352 M * daniel_hozac which is what you want for fair sharing.
1182969455 M * arekm ok but I still need to set fill/int to 2% to get hard limit (and fill2/int2 to get 2% for idle scheduling) - correct?
1182969491 M * daniel_hozac sure.
1182969616 M * arekm new question arises, is 2% of time in idle scheduling a hard limit? let say that 40% of cpu time is idle, now each guest will get only 2% of that idle time?
1182969686 M * daniel_hozac no, if there's just one guest running, it'll get all of the idle time.
1182969687 M * arekm I should be more specific. 40% is idle and there is only one guest that wants more and more power. will it get 2% of idle and never more?
1182969691 M * arekm aha, ok :)
1182969837 M * matti :)
1182969935 J * Aiken ~james@ppp121-45-220-241.lns2.bne1.internode.on.net
1182970492 N * _mcp mcp
1182970764 J * ]hd[reddot ~hdreddot@216-164-132-88.c3-0.upd-ubr8.trpr-upd.pa.cable.rcn.com
1182970767 M * slack101 what does vseerver require ?
1182970810 A * arekm needs to do some testing because it looks like guessing initial fill-rate/interval isn't easy (to get responsive system)
1182970836 P * ]hd[reddot 
1182970846 J * sauron ~zeus16384@12.20.4.100
1182970879 M * sauron now i installed 2.6.21! just I get some weird error when I try to startup a vserver
1182970900 M * sauron dynamic contexts disabled
1182970902 M * sauron then
1182970913 M * daniel_hozac so, set a static context for your guest?
1182970921 M * sauron ncontext: vc_net_create: invalid argument
1182970931 M * daniel_hozac dynamic contexts have been deprecated for years.
1182970963 M * sauron mmm. that's something new for me. I've never set any contexts yet. so I need to manually setup the xids somehow?
1182971019 M * sauron i'll read...
1182971034 J * coderanger_ ~laptop@wireless-19-114.media.mit.edu
1182971073 M * daniel_hozac it's in the FAQ...
1182971171 M * sauron works!
1182971190 M * sauron oh ok. I found it in some mail list posting.
1182971253 M * sauron wow
1182971259 M * sauron i actually used 49164
1182971271 M * sauron and faq says is must be [2-49151]
1182971274 M * sauron but it worked...
1182971308 M * daniel_hozac with dynamic contexts disabled, you can also use the 49152-65535 range.
1182971356 M * sauron great. not that I insist on it just got used to the 49... numbers.
1182971489 M * sauron i hope now the rss limit will work.....
1182971551 M * sauron YES!!! says killing process vi...
1182971556 M * sauron really cool!
1182971617 Q * lilalinux_ Remote host closed the connection
1182971619 M * arekm daniel_hozac: btw. why fair sharing between vservers isn't default option? 
1182971671 M * daniel_hozac because Linux already knows how to share between processes?
1182971705 M * arekm exactly, processes and not guests
1182971706 M * daniel_hozac defaults should be barebones, not a complete configuration for some purpose.
1182971738 M * daniel_hozac not everyone wants it to be between guests.
1182971769 M * arekm ok
1182971776 M * daniel_hozac it's much better to have users set things they want rather than unsetting a bunch of shit they don't need.
1182972540 M * arekm last thing that left for me. disks limits. I have shared xfs partition mounted at /vservers/ and I want to use it for guests but with limiting size of each guest to some value.
1182972560 M * arekm I need to use vroot devices right?
1182972584 M * daniel_hozac vroot is for user/group quotas, when you have one filesystem per guest.
1182972596 M * daniel_hozac i don't think XFS supports the disk limits yet though.
1182972677 M * arekm uh, so it seems that the only option is lvm and separate volumes  for each guest
1182972706 M * daniel_hozac for XFS, yes.
1182972717 M * daniel_hozac or if you want user/group quotas inside the guests.
1182973029 M * arekm suppose that I use ext3. Now I guess when copying inside of context then everything will be counted properly. Will something break badly if I copy something into vserver but by doing that from host-system (beside wrong counting) ?
1182973030 Q * lylix Read error: Connection reset by peer
1182973135 M * arekm maybe solution is doing x.iso and mounting as /vservers/guest/ :>
1182973517 Q * sauron Quit: bye thanks
1182973806 M * daniel_hozac nothing will break if you copy stuff from the host.
1182973827 M * daniel_hozac the count will also still be correct, since the host owns the file.
1182973935 M * arekm and host knows that I copied that file into guest X? (I'm still thinking about limiting size of hdvY as seen by guest, not typical user/grup quota)
1182973961 M * daniel_hozac hmm?
1182974106 M * arekm I have xfs on /vserver 100GB. I'm looking for a way to tell guest1 (that uses /vserver/guest) that it has "virtual hdd" of size 10GB , guest2 has "virtual hdd" of size 25GB etc
1182974121 M * arekm s/xfs/ext3/ or any other supported fs
1182974145 M * arekm so from inside of guest there is only 10GB of space visible
1182974175 Q * DavidS|Vechta Quit: Leaving.
1182974323 M * onox daniel_hozac: is there a way to limit the size of a guest?
1182974474 M * bXi installing a vserver kernel on a amd64 box now
1182974890 P * James1 
1182975555 J * lylix ~eric@dynamic-acs-24-154-33-109.zoominternet.net
1182976419 N * DoberMann[PullA] DoberMann
1182977562 M * slack101 whats a good host distro
1182977565 M * slack101 using debian 
1182977590 M * slack101 but i think i might use debian 
1182977676 Q * meandtheshell Quit: Leaving.
1182977804 N * DoberMann DoberMann[ZZZzzz]
1182977874 J * dna ~naucki@p54BCF450.dip.t-dialin.net
1182978220 Q * FireEgl Quit: Bye...
1182978969 Q * HeinMueck Quit: Aah!
1182979278 J * edeschen ~edeschen@proxy-sjc-1.cisco.com
1182979302 J * ema ~ema@rtfm.galliera.it
1182979694 Q * dna Quit: Verlassend
1182979967 J * dreamind ~dreamind@C2107.campino.wh.tu-darmstadt.de
1182979996 N * dreamind Guest101
1182982244 M * lylix daniel_hozac wasnt kidding... :)
1182983681 Q * ema Quit: leaving
1182983681 Q * Aiken Read error: Connection reset by peer
1182983731 J * Aiken ~james@ppp121-45-220-241.lns2.bne1.internode.on.net
1182983754 Q * Aiken 
1182983769 J * Aiken ~james@121.45.220.241
1182983956 Q * bonbons Quit: Leaving
1182987263 N * Bertl_oO Bertl
1182987266 M * Bertl back now ...
1182987562 M * slack101 oh shoot
1182987565 M * slack101 hes baaaaaaaack
1182987583 M * Bertl hehe :)
1182987596 M * slack101 im kinda happy 
1182987603 M * slack101 stuff is working
1182987603 M * Bertl good to hear
1182987615 M * slack101 got about 15 people back on the vserver 
1182987639 M * slack101 i wonder whats the max people ever put on one vserver 
1182987655 M * Bertl guest or host?
1182987689 M * Bertl I know Lycos tested host systems with up to 200 guests
1182987773 J * FireEgl FireEgl@FireEgl.CJB.Net
1182987909 J * rgl_ ~Rui@84.90.10.107
1182988341 Q * rgl Ping timeout: 480 seconds