1113869116 Q * kjo Ping timeout: 480 seconds
1113869443 M * mugwump romke: ok, you need to create a file /etc/vservers/foo/schedule, with *6* lines;
1113869455 M * mugwump eg:
1113869458 M * mugwump 1
1113869459 M * mugwump 20
1113869460 M * mugwump 0
1113869468 M * mugwump 0
1113869471 M * mugwump 1000
1113869473 M * mugwump dummy
1113869486 M * mugwump (tested with 0.30.196)
1113869517 M * mugwump the first line is the fill rate, the second the fill interval, giving this vserver 1/20th of one CPU allocation
1113869537 M * romke aha
1113869546 M * romke mugwump: and next lines?
1113869552 M * mugwump the third line tells it how many "tokens" to give this server when it's created.  This example gives it nothing, which is pretty nasty but the vserver will just have to cope
1113869575 M * mugwump (ie, wait 20 jiffies to be given a new token)
1113869590 M * mugwump then, the last two lines set the minimum/maximum size of the token bucket
1113869617 M * mugwump make bigger buckets for more generous bursting/smooth operation, smaller buckets to enforce allocations more quickly
1113869674 M * mugwump If you set the allocation strategy to "hard", then the vserver won't get any timeslices at all when they don't have any tokens
1113869690 M * mugwump rather than just being nice'd down to virtually nothing
1113869730 M * mugwump (btw, "don't have any tokens" is relative here - it's actually the "minimum" token bucket value)
1113869743 M * romke mugwump: fourth is what?
1113869763 M * mugwump fourth is minimum number of tokens, fifth is maximum number (size of bucket)
1113869793 M * romke mugwump: ok, and dummy? reserved? must exist?
1113869805 Q * ndim Ping timeout: 480 seconds
1113869806 M * mugwump yes, you might not need it on a newer util-vserver
1113869921 M * mugwump you can see the results of your labour by looking at /proc/virtual/NNN/sched when your vserver is running
1113869929 M * mugwump to change a running vserver you can use `vsched'
1113869995 M * romke mugwump: so: each <2> jiffies <1> tokens will be given into bucket with <5> maximum, processes will be back from "on hold" when in bucket is more than <4> - right?
1113870044 M * mugwump s/more than <4>/less than <4>
1113870058 M * mugwump sure, and they start off with <3> tokens
1113870086 J * ndim hun@helena.bawue.de
1113870206 M * romke mugwump: from Flower page: If the bucket is empty the process is put in the hold queue. When the bucket has been refilled to at least M tokens, all on hold processes are rescheduled.
1113870248 M * romke mugwump: you are sure with that "less than <4>" ?
1113870249 M * mugwump ah, right.  it's a schmidt trigger for going on hold.
1113870293 M * mugwump When I wrote it, I just made it when the bucket was depleted.  I guess that the minimum is there to stop vservers "flipping" from on-hold to running
1113870355 M * mugwump mind you, I didn't design the "hard" scheduling part at all, contexts would just "run on empty" at a very low prio
1113870554 M * romke mugwump: so basically when fill rate will be very low contexts will "work slower" than those with higher fill rate - yes?
1113870607 M * mugwump yes, assuming they all have the same fill interval
1113870613 M * romke mugwump: any recomendations on fill interval?
1113870629 M * romke mugwump: yes, of course with that assumption
1113870659 M * mugwump well, you should aim that the sum of (fill rate / fill interval) adds up to roughly the number of processors in your system
1113870706 M * mugwump ie, if you have four vservers and want them each to get an equal slice of the system, then you'd set each to (1/4)
1113870741 M * romke mugwump: yes, I know that, if it had been >> than N of processors then buckets will allways be full
1113870780 M * mugwump assuming each context always has work to do
1113870801 M * romke mugwump: but too small fill interval won't cause unnecesarry overhead? 
1113870833 M * mugwump If the fill interval is too long, it might make performance erratic.  Setting it as small as 1 or 2 is fine.  It only actually recalculates the amount when it has to
1113870862 M * mugwump of course, with a fill interval of such a small value you're dishing out large chunks of the system :)
1113870955 M * romke mugwump: "dishing out"? (explain, I don't understand this)
1113870978 M * mugwump it means (in this case) "allocating"
1113871013 M * romke ok
1113871127 M * romke mugwump: so, wouldn't be better if fill interval be more like 10-20 ? and fill rate proportionally greater?
1113871208 M * mugwump it makes little difference.  Just don't go expecting smooth scheduling behaviour when it gets near HZ (100 or 1000)
1113871261 M * romke mugwump: ok, thx for help
1113871267 M * mugwump np
1113871644 M * romke mugwump: i've just put it here: http://linux-vserver.org/Scheduler+Parameters
1113871999 M * mugwump romke++
1113872012 M * romke ++?
1113872052 M * mugwump er, it's a #perl thing.  There's a bot in there that records "karma" on things when people say (something)++
1113872138 M * romke lol
1113872248 M * romke this page should be linked from Vserver+Paper, but I dont remember on which part is scheduler explained ( I thou remeber that there is "See Example [XX]" with no example )
1113872280 A * romke > /dev/bed and maybe tommorow I'll link it...
1113872701 M * mugwump hmm, can't put unicode on linux-vserver.org wiki...
1113873677 Q * sladen Ping timeout: 480 seconds
1113874622 J * sladen paul@starsky.19inch.net
1113874673 J * ciphernaut ~a@61.88.18.130
1113875611 Q * nayco Quit: Bonne nuit !
1113878258 P * monrad Leaving
1113880711 Q * kevinp|gone Ping timeout: 480 seconds
1113880882 J * kevinp ~kevinp@ny.webpipe.net
1113881406 Q * kevinp Ping timeout: 480 seconds
1113883892 J * kevinp ~kevinp@ny.webpipe.net
1113885380 J * monrad ~monrad@213083190130.sonofon.dk
1113889116 J * MrX freebsd@219.95.232.119
1113891051 M * eyck can anyone recommend 'Testing Linux Kernels with QEMU' tutorials for dummies(me) ?
1113891537 Q * monrad Read error: Operation timed out
1113897852 J * prae ~prae@134.106-14-84.ripe.coltfrance.com
1113898662 J * FEN_HIN ~JFOC@latarius.tkdgroup.com
1113898776 Q * BWare Quit: using sirc version 2.211+KSIRC/1.3.11
1113898790 J * BWare ~bware@office.intouch.net
1113899121 Q * FEN_HIN Quit: JFOC Network v15 Professional Edition[FEN_HIN]: [- http://www.jfoc.net -] OFFLINE After 8mins 24secs off
1113899421 Q * knoppix_ Ping timeout: 480 seconds
1113899952 J * knoppix_ ~knoppix@dsl-082-082-081-066.arcor-ip.net
1113901065 J * HoxCa ~hugh@anode.atosc.org
1113901071 M * HoxCa ehlo there
1113903864 M * HoxCa got a problem with a ngnet test configuration
1113903953 M * HoxCa i got a vserver in ctx 12
1113903976 M * HoxCa cat /proc/vnet in my gentoo host seems good
1113904005 M * HoxCa and iptables -t vnet chain is looking good too
1113904142 M * HoxCa the howto said 'It's now the work of the vserver to setup network devices'
1113904220 M * HoxCa but ifconfig failed in the vserver
1113904317 M * HoxCa even if i see en0 and lo in the 'cat /proc/net/dev' inside the vserver
1113904322 J * hellekin ~hellekin@vol75-1-81-57-78-177.fbx.proxad.net
1113904328 M * hellekin aile-eau
1113904345 M * HoxCa hell eau :)
1113904352 M * hellekin =8)
1113904431 M * HoxCa anyhelp for this ngnet conf 'ifconfig failed in the vserver' ?
1113904501 M * hellekin can someone explain me the difference between newvserver, vserver build and vserver-build? On Debian I'm trying to build vservers with the new configuration framework but I only can use newvserver which builds the old /etc/vservers/host.conf instead of the directory structure... Confusion arises
1113904590 M * HoxCa hellekin: i m not on the debian side at this time
1113904628 M * hellekin i suppose i'm not on your side either as i don't know what ngnet is ;o)
1113904659 M * HoxCa but i use 'vserver build' to get my new style conf in /etc/vservers
1113905186 M * hellekin oh... I just retried here and it seems to work. But I had to disable autofs for a while
1113905222 M * hellekin (as the vserver scripts try to create dirs in the VSERVER_ROOTDIR which autofs doesn't like
1113905759 M * romke will vs2.0 contain stable ngnet ?
1113906863 M * DaPhreak nope ...
1113908033 Q * HoxCa Remote host closed the connection
1113908116 J * Linux-Life ~Odin@62.116.121.242
1113908141 P * Linux-Life 
1113908179 J * HoxCa ~hugh@anode.atosc.org
1113908188 M * HoxCa re
1113908416 N * Bertl_zZ Bertl
1113908426 M * Bertl morning folks!
1113908455 M * Bertl romke: no, 2.0 will not contain ngnet ...
1113908493 M * Bertl HoxCa: did you give the necessary Caps to configure the networking?
1113908530 M * HoxCa humm, a bcap 
1113908586 M * HoxCa i think i didnt give it
1113908591 M * Bertl CAP_NET_ADMIN
1113908598 M * DaCa hi Bertl 
1113908631 M * Bertl hey DaCa! btw, does 'Ca' mean anything I should know?
1113908645 M * Bertl (just looking at DaCa, HoxCa ;)
1113908656 M * HoxCa ;)
1113908732 M * DaCa maybe HoxCa's name starts also with Ca ?
1113908770 M * Bertl probably ... and what about 'Da'?
1113908775 M * HoxCa looks better Bertl, thank you
1113908785 M * Bertl (now looking at DaCa, DaPhreak now ;)
1113908809 M * Bertl HoxCa: you're welcome! btw, isn't that mentioned on the ngnet howto?
1113908844 M * HoxCa TODO: add caps
1113908864 M * DaCa Bertl: Da are the first 2 letters of my first name (/whois)
1113908923 M * Bertl okay, Danny, you made a perfect point.
1113908988 M * DaCa Bertl: btw, I have a question, does the chroot barrier on vserver enabled kernels (2.4/stable) work for normal tradional chroot's too? or only for vservers?
1113909027 M * Bertl well, basically it would work, but it requires the 'is inside context' to be activated
1113909091 M * hellekin Hi Bertl, long time no see
1113909100 M * Bertl hey hellekin! how are you?
1113909106 M * hellekin fine =8)
1113909111 M * DaCa Bertl: ok, so if I chcontext a daemon, which chroots itself, it should be protected?
1113909126 M * hellekin I just upgraded my kernel from 2.4.27 to 2.6.11
1113909140 M * Bertl DaCa: yes, it should ...
1113909145 M * hellekin messing with autofs + lvm2 to setup a diskfull of vservers
1113909150 M * Bertl hellekin: not bad ...
1113909171 M * Bertl what issues has autofs(4?)
1113909216 M * hellekin if i put: /var/lib/vservers :/etc/auto.vservers in /etc/auto.master
1113909229 J * IceTi 125@mw635449.rz.ruhr-uni-bochum.de
1113909237 M * hellekin then try to install a vserver, it won't as autofs forbids create new dirs
1113909252 M * hellekin i have to disable autofs in order to build a new vserver
1113909270 M * IceTi hi
1113909272 M * hellekin either i'm missing something about autofs (which i surely do)
1113909279 M * Bertl hey Timo!
1113909284 M * IceTi :-)
1113909306 M * hellekin or the install scripts should be modified to load their tmp, .defaults and .pkg dirs 
1113909314 M * IceTi In which file can I type the packages that I want to install when I make vserver build ???
1113909321 M * hellekin in e.g. /etc
1113909336 M * hellekin argh... gotta run (and eat)... cul8r
1113909353 M * Bertl k, cya
1113909498 M * Bertl IceTi: there is a default file (described on the flower page) IIRC
1113909514 M * IceTi yes I search in the flower page yet
1113909666 M * IceTi but where is this file?
1113909671 Q * prae Ping timeout: 480 seconds
1113909674 M * IceTi i really don´t find that :-(
1113909679 M * IceTi and I have searches
1113909685 J * prae ~prae@134.106-14-84.ripe.coltfrance.com
1113909685 M * IceTi searched
1113909758 M * Bertl well, looking at the flower page it says:
1113909770 M * Bertl /etc/vservers/.distributions 
1113909788 M * Bertl /etc/vservers/.distributions/dist/pkgs 
1113909794 M * Bertl Contains files with packages.
1113909794 M * Bertl     * list
1113909794 M * Bertl       File which contains the name of packages. On top of file the special keywords '--reinstall' and '--can-fail' are possible.
1113909836 M * IceTi hmm /dist/pkdgs --> that directory doesn´t exist
1113909866 M * Bertl well, you should really read the page again ... 
1113909902 M * IceTi I don´t wanna say somethinf about this page but its not soo good
1113909904 M * IceTi ;-)
1113909923 M * Bertl go ahead, make a better one ;)
1113909934 M * IceTi yes?
1113909934 M * HoxCa still non network cannot ping my host server from the vserver. iptables -t vnet -L looks good; any hints ?
1113909936 M * IceTi really?
1113909947 M * IceTi some context but another style?
1113909957 M * IceTi if I ahve time i will do it ;-)
1113909958 M * Bertl HoxCa: for now, arp is not implemented, you need to add static arp entries
1113909968 M * HoxCa humm :)
1113909979 M * Bertl IceTi: you can select the style, if your browser is able to do so ;)
1113909985 M * HoxCa yes! i will sir!
1113909991 M * IceTi hmm
1113910015 M * romke Bertl: can I add page http://linux-vserver.org/Scheduler+Parameters to Linux-VServer-Paper-06 (as example?)
1113910050 M * IceTi must I amke the file pkgs ??
1113910073 Q * hellekin Quit: No windows for this server
1113910085 M * Bertl romke: hmm, I'd prefer a cleanup first ...
1113910137 M * Bertl romke: a few suggestions here:
1113910173 M * Bertl don't use the <1> <2>, IMHO it's too confusing, better use <fillRate> or emphasisze it with italics or so
1113910242 M * Bertl also some conclusions sound a little far fetched, I'd like to hear a rationale to that ...
1113910348 M * IceTi or is pkdgs a file?
1113910386 M * IceTi or a directory and then a file name list ???
1113910423 M * Bertl with so many ors it's easy to answer: yes ;)
1113910473 M * IceTi hehe and what is it now?
1113910529 M * Bertl well, did you look at the existing distros on your system?
1113910536 M * IceTi yes
1113910543 M * Bertl and, what did you see there?
1113910556 M * IceTi Iám in fc3 now
1113910572 M * Bertl so is there a file or a dir named pkgs ?
1113910579 M * IceTi no
1113910583 M * IceTi i have make this file
1113910590 M * IceTi touch pkgs
1113910606 M * romke Bertl: fixed <1> <2>..., conclusions are pasted from discussion with mugwump (/me knows to little to make such)
1113910652 M * Hollow hey Bertl 
1113910657 M * Bertl romke: okay, so I'd say, link it from the Documentation page ...
1113910700 M * Bertl (because I'm not convinced of some conclusions there, but without more investigations I can not verify/falsify them easily)
1113910711 M * Bertl hey HoxCa!
1113910725 A * Bertl *arg*
1113910732 M * Bertl hey Hollow! ;)
1113910755 A * Bertl is getting lazy ;)
1113910787 M * Hollow i tried to get the gentoo init system running with init inside a vserver, it boots fine, but everytime i start a vserver my keyboard is gone... everything runs fine, mouse, X, etc, but no keyboard input anymore..
1113910791 M * Hollow anything known?
1113910821 M * Hollow init style was plain
1113910829 M * Bertl no idea ... but X is fine sounds like you give a lot of permissions, no?
1113910850 M * Bertl or do you mean, X on the host?
1113910852 M * Hollow to the vserver? the standard ones
1113910855 M * Hollow yep
1113910859 M * HoxCa i use init style gentoo, but i didnt use X ;)
1113910888 M * IceTi :-(
1113910892 M * IceTi what is now?
1113910917 M * Bertl Hollow: could you upload the scripts your init executes on startup?
1113910919 M * IceTi a file name pkdgs? or a directory ?
1113910965 M * Hollow Bertl: yup, will take a bit since i completely changed it already to boot with gentoo init style again, but i will revert them
1113910977 M * Bertl IceTi: I would go for a dir, because it's hard to put a file in a file yet ;)
1113910997 M * IceTi haha ;-)
1113911003 M * IceTi and then a fiel name list?
1113911018 M * IceTi and when i make build this will catch this file?
1113911021 M * Bertl okay, back after lunch break ...
1113911028 N * Bertl Bertl_oO
1113911773 J * DuckKing ~Duck@dyn-83-157-129-150.ppp.tiscali.fr
1113911941 J * hellekin ~hellekin@vol75-1-81-57-78-177.fbx.proxad.net
1113911946 M * hellekin re
1113912009 Q * DuckMaster Read error: Operation timed out
1113912753 M * DaPhreak damn bubbly ...
1113913578 J * ozan ~ozan@85.96.199.40
1113913765 M * ozan hi all, some body runed asterisk in vserver? with iax trunking enabled?
1113914711 Q * HoxCa Quit: Garfield?s philosophy?
1113915179 J * micah_ micah@micha.hampshire.edu
1113915240 Q * ozan Quit: User abort with 5 Ctrl-C's
1113915598 Q * micah Ping timeout: 480 seconds
1113916281 M * wurd hi
1113916293 M * wurd how come i dont even have the "passwd" command inside my vserver ?
1113916311 M * wurd also, my prompt says "bash 3.00", is this normal ?
1113916381 Q * IceTi Quit: get satisfied! • :: ««« (Gamers.IRC) »»» www.gamersirc.net ::
1113916477 Q * Hollow Remote host closed the connection
1113916522 J * Hollow ~Hollow@home.xnull.de
1113916713 M * wurd hi hollow, could you give me a hand ?
1113916719 M * Hollow sure
1113916787 M * wurd i'd like to know why i dont even have the "passwd" command inside my vserver
1113916790 M * wurd also, my prompt says "bash 3.00", is this normal ?
1113916959 J * micah micah@micha.hampshire.edu
1113917080 Q * micah_ Ping timeout: 480 seconds
1113917085 M * Hollow wurd: the passwd command is provided by the shadow package
1113917122 M * Hollow if you bash shows the standard prompt it seems it hasn't loaded /etc/profile or so
1113917228 M * wurd okay so i need the shadow package....
1113917235 M * wurd and tons of other packages also....
1113917240 M * wurd what should i do ?
1113917257 M * wurd is this normal that i lack of so many packages ?
1113917344 M * Hollow dunno, your setup seems quite strange to me ;)
1113917374 M * wurd what should i know that i dont know
1113917396 M * wurd i find it strange also, since it lacks so many packages
1113917406 M * Hollow how did you setup the vserver?
1113917420 M * wurd i built it, thats it
1113917424 M * Hollow how?
1113917466 M * wurd vserver servername build -m apt-rpm --hostname test --interface eth0:blab.bla.bla/bla -- -d fc3
1113917492 M * Hollow mhm.. seems you have fedora core 3
1113917496 M * wurd yes
1113917539 M * Hollow i don't know what the apt-rpm built method sets up, but it seems you lack some packages... doesn't fedora have a nice pkg manager? ;)
1113917553 N * Bertl_oO Bertl
1113917560 M * Hollow wb Bertl 
1113917567 M * Bertl thanks!
1113917578 M * Bertl Hollow: yes it has, it's called apt-rpm ;)
1113917593 M * Bertl wurd: which tools did you use/try?
1113917594 M * Hollow Bertl: this is the baselayout used in the vsrver: http://home.xnull.de/work/gentoo/vserver/baselayout/baselayout-vserver-1.0.tar.bz2
1113917666 M * Hollow Bertl: after restarting X (over ssh) the keboard worked again
1113917681 M * Hollow maybe a X/KDE problem?
1113917709 M * BWare Hollow: are you using an acpi based system, or a 'touchpad' driver ?
1113917735 M * BWare and a 2.6 kernel
1113917739 M * Hollow seems so, yes
1113917758 M * BWare run a dmesg and check for syn messages
1113917765 M * BWare sync even
1113917791 M * Hollow apropos dmesg... my dmesg is full of http://phpfi.com/58537
1113917798 M * Hollow looks like vserver debugging code
1113917825 M * BWare looks like it ;)
1113917849 M * Hollow but no sync
1113917853 M * Bertl hmm .. hmm ...
1113917855 M * BWare all I know is from my experience with 2.6 on an acer laptop and the touchpad driver is that it hangs
1113917861 M * Bertl Hollow: what arch do you use?
1113917866 M * Hollow i386
1113917874 M * BWare and dmeg shows a lot of lost sync messages
1113917876 M * Bertl and what tools are those?
1113917881 M * Hollow 205
1113917898 M * wurd <Bertl> wurd: which tools did you use/try?
1113917898 M * Bertl why would they use xid=65535?
1113917902 M * wurd what do you mean ?
1113917905 M * BWare So I gave up the touchpad and started using imps/2 once again
1113917907 A * Hollow shrugs
1113917927 M * Bertl wurd: 0.30.204, 0.30.205 or 0.30.206?
1113917930 Q * hellekin Quit: No windows for this server
1113917931 M * wurd 205
1113917933 J * IceTi 125@mw635449.rz.ruhr-uni-bochum.de
1113917989 M * Bertl Hollow: please do 'echo 1 >/proc/sys/vserver/debug_switch'
1113918010 M * Bertl then start the server once, and upload the debug output in klog
1113918077 M * Hollow k
1113918278 M * wurd Bertl why did you ask?
1113918339 M * Bertl because fc3 was not really supproted before 0.30.205
1113918384 Q * Hollow Remote host closed the connection
1113918414 J * Hollow ~Hollow@home.xnull.de
1113918452 Q * IceTi Quit: get satisfied! • :: ««« (Gamers.IRC) »»» www.gamersirc.net ::
1113918476 M * Hollow Bertl: http://phpfi.com/58538
1113918984 M * wurd do you know why i lack so many packages?
1113919038 M * wurd did i do something wrong
1113919297 M * Bertl wurd: no idea, you ahve to ask enrico for details ...
1113919433 M * Bertl wurd: didn't see any posting on the ML yet ... so you might try that ...
1113919597 M * wurd so it *ISNT* normal that i lack so many packages ?
1113919619 M * wurd normally when you enter a freshly-built vserver, it is full of all kinds of packages ?
1113919643 M * SiD3WiNDR I think and hope it would be as minimal as possible
1113919667 M * wurd i find it a bit "too mininal" since i dont even have the "rpm" command
1113919674 M * wurd nor "passwd"
1113919695 M * wurd but i'd like to know if this is NORMAL.
1113919717 M * wurd or if it is that way because i'm experiencing problems.
1113919733 M * DaCa looks a bit too minimal indeed :)
1113919868 M * DaPhreak only a bit DaCa ?! :)
1113920072 M * wurd okay, so i DO have problems.
1113920076 M * wurd at least i know that now
1113920465 M * DaCa DaPhreak: they could have least dropped a busybox in it :)
1113920497 M * DaCa at least even
1113922021 M * Bertl wurd: what packages _are_ installed there?
1113922270 Q * ripman Ping timeout: 480 seconds
1113922965 M * kevinp morning all!
1113923039 M * kevinp Bertl were you looking for me a few days ago?
1113923073 M * DaPhreak hehe yeah but who needs busybox inside a vserver DaCa .. you wouldn't run a vserver on a embedded device with 1GB flash-ram ;) 
1113923098 M * kevinp wurd: your experience is the same as mine, it only installs a few packages - haven't we already gone through this once?
1113923205 M * kevinp 1113848725 M * kevinp wurd: you also used the option -m apt-rpm which means that you installed the rpms specified in /usr/lib/util-vserver/distributions/fc3/pkgs which is very little unless you added more to it
1113923220 M * Bertl kevinp: probably ;)
1113923230 M * wurd so you're saying i need to download a whole bunch of packages and putting them in the folder ?
1113923252 M * wurd like, hundreds of them
1113923253 M * Bertl wurd: normally apt-get is supposed to get a bunch of packages for you ...
1113923302 M * kevinp 1113849051 M * kevinp wurd: you can add more packages to the 01, 02, and 03 files in the pkgs folder I mentioned before and create another vserver with all of the packages at once
1113923317 M * kevinp 1113849128 M * kevinp for example, when I want to a minimal install of centos in a vserver, I use all of the packages here: http://linux-vserver.org/CentOS_min_install
1113923319 M * wurd Bertl how can i know which packages were installed? i dont even have the 'rpm' command :)
1113923369 M * DaPhreak cu tomorrow guys :) gonna do some more partying at home ;)
1113923379 M * kevinp wurd: the pkgs folder has several files in it that are a list of all of the rpms that you would like to download and install when you create a new vserver
1113923392 N * DaPhreak DaPhreak|screen_detached
1113923405 M * Bertl wurd: try vrpm
1113923423 M * wurd Bertl , nope
1113923434 M * wurd kevinp , cool but how do i know which packages to put in there?
1113923439 M * wurd i probably need hundreds of them
1113923458 M * kevinp Like I said last time, you can start by looking what you have installed on your host server
1113923474 M * kevinp rpm -qa > mypkgs
1113923495 M * kevinp then remove the ones you really don't want in your vserver
1113923563 M * wurd good idea, thanks
1113923584 M * kevinp If you check with the Fedora people there is probably a comp.xml file that you can download that lists what rpms are included with each checkbox on an install
1113923592 M * wurd then, i rebuild the vserver  (vserver servname build -m apt-rpm etc etc..) and thats it ?
1113923607 M * kevinp well that's a good beginning
1113923674 M * kevinp Bertl: did you see my messages a few days ago that klogd does not start still with 2.0pre1?
1113923782 M * Bertl wurd: what does nope mean?
1113923809 M * Bertl let me rephrase: what do you mean by nope?
1113923816 M * wurd lol.
1113923831 M * wurd it means that i dont have the 'vrpm' command :)
1113923839 M * Bertl ahem?
1113923849 M * Bertl it's part of the tools you claim to have installed?
1113923863 M * wurd hmm
1113923886 M * wurd is there a way i can verify they have indeed been installed correctly?
1113923900 M * Bertl rpm -V ?
1113923926 M * Bertl kevinp: could you strace the klogd with 2.0pre1?
1113923943 M * wurd util-vserver-build-0.30.205-0
1113923943 M * wurd util-vserver-devel-0.30.205-0
1113923943 M * wurd util-vserver-lib-0.30.205-0
1113923943 M * wurd util-vserver-debuginfo-0.30.205-0
1113923950 M * wurd util-vserver-legacy-0.30.205-0
1113923950 M * wurd util-vserver-0.30.205-0
1113923950 M * wurd util-vserver-sysv-0.30.205-0
1113923950 M * wurd util-vserver-core-0.30.205-0
1113923988 M * wurd ah i get it man, sorry
1113923997 M * wurd i thought you wanted me to type 'vrpm' inside my vserver
1113923999 M * wurd :/
1113924023 M * Bertl hmm, oaky ;)
1113924503 J * Jd0gg Jdogg@c-24-1-174-133.hsd1.tx.comcast.net
1113924509 Q * Jdogg Read error: Connection reset by peer
1113924542 N * Jd0gg Jdogg
1113924638 M * prae DaCa: here my dear ?
1113924680 M * DaCa prae: whatsup?
1113924697 M * prae DaCa: have you merged vs with grsec ?
1113924754 M * DaCa prae: http://vserver.limehouse.org # it's on the wiki too
1113924767 M * prae oh !
1113924769 M * prae hmmm
1113924771 M * prae :)
1113924897 M * prae DaCa: hmmmm, good  :)
1113924900 M * prae DaCa: thanks  =)
1113924967 M * wurd kevinp why is there three files (01,02,03) ?
1113925009 M * DaCa prae: let me know if it works ok, I use 2.4 myself and have only loosely tested the 2.6 patch (that means I compiled it and ran testme.sh :))
1113925042 M * prae :)
1113925056 M * prae hmmm, I test now into 2U
1113925085 M * prae SMP: (2 procs) one with HT, one without HT
1113925153 M * prae I send you the feedback after :)
1113925487 M * wurd Bertl 
1113925491 M * wurd are you still there
1113925511 M * Bertl yup, I am?
1113925561 M * wurd wait, i think i answered my own question - thanks anyway :)
1113925922 M * romke where did my /prov/virtual go? (may it be because of grsec?)
1113925975 M * romke s/v/c/
1113926117 M * Bertl could be, no idea ...
1113926191 M * DaCa Bertl: is there supposed to be a /proc/virtual in 1.2.10?
1113926243 M * wurd is there another way to build a vserver than with -m apt-rpm
1113926243 M * wurd  ?
1113926274 M * Bertl yes, plenty of ...
1113926298 M * romke hmm, i'll try without grsec then...
1113926322 J * erwan_taf ~erwan@AToulouse-105-2-2-69.w217-128.abo.wanadoo.fr
1113926345 M * Bertl wurd: you can use debootstrap, or list based rpm, or existing images/templates or even your own build/install command
1113926509 M * wurd which one will provide my vserver with "not too minimal" setup ?
1113926536 M * wurd i.e will include the RPM, PASSWD, VIM commands, to name a few . ..
1113926593 M * Bertl all of them including the apt-rpm based ;)
1113926604 M * wurd the way apt-rpm doesnt seem to work
1113926607 M * Bertl wurd: I'm pretty sure something is going wrong ...
1113926610 M * wurd Reading Package Lists... Done
1113926610 M * wurd Building Dependency Tree... Done
1113926612 M * wurd E: Couldn't find package device-mapper-1.00.19-2
1113926617 M * Bertl ah, see?
1113926639 M * wurd well?
1113926647 M * Bertl did you select an apt rpm source and verify that it is up to date?
1113926713 M * wurd this you mean? /etc/apt/sources.list
1113926822 M * Bertl no, the one inside /etc/vservers...
1113926864 M * romke Bertl: yep, its grsec 2.1.4 (removing /proc/virtual)
1113926894 M * Bertl well, the real advantage of grsec ;)
1113926985 M * wurd this? /vservers/.pkg/vserv12/apt/etc/sources.list
1113927032 M * romke Bertl: lol
1113927072 M * Bertl wurd: you should have one in *distrib*
1113927086 M * Bertl sec, looking for the full path
1113927101 M * Bertl /etc/vservers/.distributions/fc3/
1113927140 M * wurd okay theres a bunch of commented lines
1113927143 M * wurd and one thats uncommented
1113927173 M * Bertl okay, try to comment that and get another uncommented
1113927198 M * Bertl then try to build a new vserver
1113927216 M * wurd yeah
1113927220 M * wurd thats what i did last time
1113927234 M * Bertl and you get the same E: blabla?
1113927234 M * wurd comment, uncomment another
1113927245 M * wurd well at first the build didnt work
1113927256 M * wurd so i did what you told me : comment, uncomment
1113927272 M * wurd then, the build worked with the 4th line uncommented
1113927275 M * wurd so i left it like that
1113927291 M * Bertl well, E: is an error, so I don't think it really worked ;)
1113927315 A * Bertl doesn't use apt-* yet, but at least I think it is ...
1113927400 M * wurd i mean..
1113927411 M * wurd the E: error i just pasted you
1113927428 M * wurd it showed up when i added a bunch of packaged in the 01, 02, 03 files
1113927432 M * wurd packages*
1113927447 M * wurd while rebuilding
1113927810 M * wurd ok, i've just tried uncommenting all the lines (one at a time) and none worked :(
1113927844 M * Bertl okay, so I probably suggest the same as last time ...
1113927880 M * Bertl please try the latest version (0.30.206) and if the issue remains, please post some kind of bug report to the mailing list, giving some examples and details ...
1113927895 M * wurd okay..
1113927897 J * kalou ~kalou@84.5.86.6
1113927900 M * kalou 'lo
1113927928 M * wurd will the 206 change the sources.list file ?
1113927933 M * Bertl wurd: I have no fc3 anywhere and I have no idea how to test it ... enrico maintains the tools and probably can test fc3 building too
1113927944 M * Bertl hey kalou!
1113927948 M * kalou :)
1113927993 M * kevinp Sorry, stepped away.  wurd: is device-mapper-1.00.19-2 available on those repositories normally?
1113928010 M * kevinp can you do 'yum search device-mapper' on your host and have it work?
1113928033 M * Bertl btw, there should be a yum install method too, IIRC
1113928065 M * kevinp that's correct, but search will tell us if it's even there
1113928090 M * Bertl no, I meant for util-vserver ;)
1113928122 M * kevinp oh, yeah, definitely, that's how I'm doing all of mine
1113928370 M * kalou I have a silly behaviour here with tcpdump inside a vserver. (2.4.26-vs1.28). The vserver has *no cap at all* but tcpdump works
1113928394 M * Hollow Bertl: any idea for my keyboard problem?
1113928411 M * Bertl kalou: hmm .. interesting ...
1113928415 M * kevinp Bertl: Did you see my question about this a few days ago? http://deadbeefbabe.org/paste/354
1113928435 M * kevinp certain rpms don't install correctly inside a vserver
1113928442 M * kalou Bertl: I restarted the vserver after removing the CAP_* in /etc/vservers
1113928448 M * kevinp others work fine
1113928452 M * Bertl Hollow: did you paste an url for the debug output?
1113928478 M * Hollow yup: http://phpfi.com/58538
1113928499 M * kalou are there any way, on 1.28, to see the active capabilities for a vserver ?
1113928551 M * Bertl /proc/self/status
1113928572 M * Bertl Hollow: hmm, so the !!! context happens without any vserver interaction?
1113928575 M * kalou CapPrm: 00000000fffffeff
1113928575 M * kalou CapEff: 00000000fffffeff
1113928575 M * kalou CapBset:        00000000fffffeff
1113928588 M * Bertl means it has _all_ capabilities ;)
1113928597 M * kalou :-(
1113928600 M * kevinp this is the output with of installing the rpm manually with full verbosity: http://deadbeefbabe.org/paste/355
1113928621 M * Hollow Bertl: all this happend with "vserver gentoo start" it has a static context 3253
1113928638 M * kalou Bertl: so empty S_CAPS variable means all capabilities ? ... grr
1113928648 Q * prae Quit: Client exiting
1113928674 M * Bertl kalou: normally not ...
1113928686 M * Bertl kalou: could you upload the output of testme.sh?
1113928696 M * Bertl Hollow: so what failed?
1113928717 M * kalou Bertl: yes, just give me the testme.sh
1113928724 M * kevinp Bertl any specific flags on the strace of klogd?
1113928734 M * Bertl kevinp: -fF
1113928750 M * Hollow Bertl: i started the vserver from a kde Konsole, and after starting X has no keyboard input anymore.. after  killing X it has keyboard input again, anything else runs fine
1113928802 M * Hollow it runs also fine with the gentoo init style, but i modified the rc-scripts to be able to run with init, and changed init style to plain
1113928810 M * Bertl Hollow: sorry, I'm trashing ;)
1113928832 M * Bertl kevinp: so what failed?
1113928851 M * kevinp on the rpm?
1113928856 M * Bertl yup
1113928871 M * kevinp error: unpacking of archive failed on file /usr/share/locale/fi/LC_MESSAGES/up2date.mo;4265322b: cpio: lsetfilecon
1113928883 M * kevinp and selinux is disabled
1113928889 M * Bertl yeah, I see that, but the last time it failed on fi, no?
1113928902 M * Bertl this time it failed on he ...
1113928920 M * Bertl so it got about 3 files further?!
1113928952 M * kevinp hmm, so keep running it until it works? :)
1113928958 A * Bertl was never good at interpreting cpio output
1113928969 M * Bertl so I would like to know _what_ failed there ;)
1113928991 M * Bertl man lsetfilecon
1113928991 M * Bertl No manual entry for lsetfilecon
1113929008 M * Bertl is this on NFS?
1113929028 M * kevinp no
1113929051 M * kevinp it does keep going further almost every time I run it
1113929062 M * kevinp it is a selinux api
1113929080 M * Bertl well, then get rid of selinux ...
1113929110 M * kevinp here's the man: http://deadbeefbabe.org/paste/356
1113929129 M * kevinp it is disabled, but the library is still tied in to a lot of things
1113929191 M * Bertl well, so what are you telling me, an selinux comman fails in cpio, but selinux is disabled?
1113929212 M * kevinp I guess... shrug
1113929214 M * Bertl to me it looks like you are asking on the wrong channel ... ;)
1113929225 M * kevinp but it works fine on the host
1113929241 M * romke i've set schedule for context 108: 2 / 20 / 100 / 40 / 200 / dummy, for context 109: 18 / 20 / 100 / 40 / 200 / dummy, ran on each burnP6 and both contexts report equal userTIME - why?
1113929288 M * Bertl kevinp: good, make an strace of a cpio run which works on the host, and a second one which fails on the guest ...
1113929348 M * Bertl romke: what flags?
1113929362 M * Bertl (cflags)
1113929631 M * romke Bertl: thx, now i know it must be sched_hard
1113929677 M * Bertl so I assume it works now?
1113929731 M * romke yeah
1113929810 M * romke Bertl: if sched_prio then limits will be executed by setting priorites - yes?
1113929842 M * Bertl prio and hard are independant
1113929854 M * Bertl prio means that the token bucket influences the priorities
1113929874 M * Bertl hard means that the token bucket suspends tasks once it is empty
1113929901 M * romke Bertl: ok, i get it now, thx
1113929909 M * Bertl you're welcome!
1113930318 M * romke Bertl: I can't stop be amazed by Vserver :>
1113930353 M * kevinp Interesting.  I did a fresh install of a new vserver, without the selinux-policy-targeted-1.17.30-2.52.1 - the library is still required due to dependencies and the rpms installed just fine
1113930637 M * Bertl romke: good so ;)
1113930667 M * Bertl kevinp: which means?
1113930814 M * romke Bertl: possible sollutions for my hosting company: reducing failover (rsync vserver image + heartbeat over com), limiting situations when one client generates load > 100 and others can just sit and watch. And it's only begining :>
1113930991 M * Bertl okay, sounds good ...#
1113931005 M * Bertl have to leave now, will be back in the evening I guess ...
1113931012 N * Bertl Bertl_oO
1113931121 M * kevinp Bertl_oO: I guess it would mean that selinux wasn't really disabled or something///
1113931204 Q * MrX Remote host closed the connection
1113931215 A * kalou is back
1113931232 J * ripman leukejong@d594e3d7.dsl.concepts.nl
1113931266 M * romke Hollow: kernel vserver-sources-1.9.5.ebuild, /proc is missing ./virtual entry. Could you take look at this issue?
1113931451 M * Hollow where is /proc missing?
1113931549 N * Doener|gone Doener
1113931553 M * romke Hollow: /proc/virtual is missing
1113931558 M * Doener evening folks!
1113931571 M * Hollow romke: on the host? do you have a vserver kernel?
1113931584 M * romke Hollow: yes, on the host
1113931619 M * Hollow do you have vserver support compiled in? ;)
1113931643 M * romke Hollow: i have 2 alternative vserver kernels, 2.6.11.6+vs1.9.5 and 2.6.11.6+vs1.9.5+grsec2.1.4
1113931688 M * romke Hollow: both with vserver support compiled in, the one with grsec have no /proc/virtual
1113931706 M * Hollow did you enable grsec?
1113931844 M * romke Hollow: yes
1113931863 M * Hollow can you try if it works with grsec disabled?
1113931948 M * wurd kevinp sorry i was away
1113931957 M * wurd i did 'yum search device-mapper'
1113931969 M * romke Hollow: so apply patches but disable it in .config?
1113931975 M * Hollow yup
1113931995 M * wurd i didnt receive any errors so i assume its working ?&
1113932144 M * romke Hollow: i'll try, unfortunatelly not today...
1113932152 M * Hollow ok
1113932831 M * kevinp wurd, so did you only get the one error and then it continued or it stopped?
1113932996 M * wurd when%
1113932996 M * wurd ?
1113933066 M * wurd when doing 'yum search device-mapper'   ?
1113933522 M * kevinp when installing a new vserver
1113933585 M * wurd <wurd> i didnt receive any errors so i assume its working 
1113933598 M * wurd this was concerning  'yum search device-mapper'
1113933618 M * wurd i have added a package list in the 01 02 03 files
1113933621 M * wurd since then,
1113933625 M * wurd when i build a new vserver
1113933636 M * wurd i get the error that 'whatever package is first in the list' cant be found
1113933656 M * romke where reports bugs about util-vserver ?
1113933704 M * romke vserver-copy should use /bin/bash instead of /bin/sh
1113934385 M * DaCa romke: you mean it shouldn't have bashims? :)
1113934419 M * DaCa romke: which one did you discover by the way?
1113935436 J * kjo nobody@p5484FD4F.dip.t-dialin.net
1113935940 M * romke DaCa: ?
1113935987 M * DaCa romke: why should vserver-copy use /bin/bash instead of /bin/sh ?
1113936018 M * romke DaCa: vserver-copy has #!/bin/sh, and in 137 line stopstart=(false) - which is bash syntax for sure
1113936081 M * romke DaCa: minimal /bin/sh may not support this syntax
1113936260 Q * Jdogg Read error: Connection reset by peer
1113936320 J * Jdogg ~My@c-24-1-174-133.hsd1.tx.comcast.net
1113936384 M * DaCa romke: indeed
1113936831 J * nayco ~nayco@82.249.51.112
1113936837 M * nayco 'llo !!!
1113936912 Q * romke Quit: [new kernel?]
1113937260 M * wurd cant find the vserver utils on the website :/ 
1113937636 M * wurd ok... found it finally 
1113937637 M * wurd :S
1113937873 Q * kjo Read error: Operation timed out
1113939718 Q * erwan_taf Read error: Operation timed out
1113939728 J * SNy_ ~mfr@bmx-chemnitz.de
1113939728 Q * SNy Read error: Connection reset by peer
1113939798 N * SNy_ SNy
1113940888 J * erwan_taf ~erwan@AToulouse-105-2-2-69.w217-128.abo.wanadoo.fr
1113941994 J * romke romke@procyon.romke.net
1113942751 Q * erwan_taf Ping timeout: 480 seconds
1113943632 N * kevinp kevinp|gone
1113943812 J * Nik ~Nik@cable-153-130.online.bg
1113943816 M * Nik hi all
1113944222 J * sOCiETY ~society@p549AD5BF.dip.t-dialin.net
1113944225 M * sOCiETY hi
1113944271 M * sOCiETY bertl you awake?
1113944306 Q * Nik Ping timeout: 480 seconds
1113944319 J * Nik ~Nik@cable-153-130.online.bg
1113944609 Q * sOCiETY Quit: Verlassend
1113944625 Q * nayco Quit: Bonne nuit !
1113944668 J * sOCiETY ~society@p549AD5BF.dip.t-dialin.net
1113944801 Q * Nik Ping timeout: 480 seconds
1113945023 Q * sOCiETY Quit: Verlassend
1113945133 J * Nik ~Nik@cable-153-130.online.bg
1113945926 Q * Nik Ping timeout: 480 seconds
1113947041 N * Bertl_oO Bertl
1113947059 M * kalou Bertl !
1113947068 M * Bertl hey kalou!
1113947092 M * kalou I found the source of my CAP_ problem on 2.4.26-vs1.28
1113947100 M * Bertl module?
1113947116 M * kalou In fact, the vserver root is a symbolic link to a directory :-(
1113947136 M * Bertl hmm?
1113947160 M * Bertl (means what has that to do with CAPs?)
1113947188 M * kalou Don't know.. but this vserver is the only one to mess with CAPs
1113947189 J * yarihm ~yarihm@80-218-3-32.dclient.hispeed.ch
1113947221 M * kalou any CAP that I set or remove, using the S_CAPS variable is ignored, and the vserver is given all the CAPs
1113947223 M * Bertl welcome yarihm!
1113947251 M * yarihm yo Bertl 
1113947261 M * Bertl kalou: doesn't manage to convince me ..
1113947284 M * Bertl could you upload the output of a vserver --debug <name> start?
1113947288 M * kalou yep
1113947289 M * yarihm hey Bertl, a question ... is there some sort of hostility between vserver and xen?
1113947306 M * Bertl no, definitely not from our side! why?
1113947326 M * yarihm just a question
1113947345 M * yarihm :) there was once some discussion in here about a paper dealing with performance-comparison
1113947427 M * Bertl ah, regarding the FSM article maybe?
1113947504 M * kalou humpf .. No configuration for this vserver: /etc/vservers/--debug.conf
1113947519 M * yarihm Bertl: yeah, maybe
1113947541 M * kalou (vserver --debug xxx start fails, vserver xxx --debug start fails too)
1113947558 M * kalou vserver-utils-0.30
1113947566 M * Bertl kalou: hmm, you probably have old tools .. yeah, right ...
1113947577 M * Bertl okay, then do the following:
1113947619 M * Bertl hmm, the vserver is a script, right?
1113947628 M * kalou yes
1113947631 J * Jani Jani@G90ba.g.pppool.de
1113947638 M * Jani hii all
1113947638 M * Bertl welcome Jani!
1113947685 M * Bertl kalou: okay, add -x to the shebang line
1113947745 M * kalou No configuration for this vserver: /etc/vservers/-x.conf
1113947747 M * Jani I have update from an old vserver version (with the 2.2er kernel) to vserver what uses 2.6 (http://linux-vserver.org/Step-by-Step+Guide+2.6), the testme.sh shows the 031 and 202 is failing. Where I get a list what the numbers are saying?
1113947769 M * kalou I can download other tools maybe ?
1113947821 M * Bertl Jani: here or in the source ...
1113947844 M * Bertl kalou: no, what I meant was: change the #!/bin/bash to #!/bin/bash -x there 
1113947912 M * Jani Bertl: 031 means that the hostname would changed to zaphod? Or?
1113947919 M * kalou hmm. silly am I, off course
1113947933 M * Bertl Jani: sec checking
1113948477 M * Bertl Jani: was a longer sec ...
1113948484 M * Jani No problem.
1113948486 M * Jani I have time :)
1113948489 M * Bertl 31 is utsname virtualization
1113948504 M * Bertl 202 is fakeinit support
1113948519 M * Bertl you are on non x86?
1113948549 M * Doener hm, dynamic context?
1113948553 M * Jani It's a x86 machine. AMD 2300+ Sempron with Debian.
1113948566 M * Doener IIRC those we're the tests that fail with unpatched .205
1113948587 M * Jani I have follow the Howtwo from http://linux-vserver.org/Step-by-Step+Guide+2.6
1113948590 M * Doener Jani: try this patch http://www.13thfloor.at/~doener/vserver/patches/diff-util-vserver-0.30.205-chcontext-dynamic-fix.diff
1113948590 M * Jani *hmms*
1113948597 M * Jani Ok
1113948599 M * Doener (against util-vserver, not the kernel ;)
1113948645 M * Doener s/we're/were/ (somewhere above ;)
1113948650 M * Bertl Doener: hmm, utsname fails?
1113948665 M * Doener no, chcontext with dynamic context fails 
1113948690 M * Jani Doener:  That I have thought as I have look into the file ;)
1113948715 M * Bertl Doener: ahh, good point, I should mark that somehow ...
1113948809 M * Doener Jani: btw, did you like the guide? any suggestions for improvement?
1113948834 Q * Loki|muh Remote host closed the connection
1113948834 M * Jani Doener: Let me first creat the first runnable vserver with it ;)
1113948840 M * Jani But sofar it's good.
1113948850 M * Doener well, that part isn't covered yet ;)
1113948859 M * Jani hehe
1113948866 M * Jani It's in the alpha docu?
1113948869 M * kalou Bertl: did you get the file ?
1113948912 M * Doener Jani: yep, that page isn't going too much into details, but for now it has to be sufficient
1113948915 J * Loki|muh loki@satanix.de
1113948922 M * Jani Do I have big problems to get my old vserver running with the new version?
1113948934 M * Doener They should run out-of-the-box
1113948950 M * Jani Doener: *nods* The HowTo is nice, and everyone who know how to handle linux should be able to create a running system
1113948950 M * Doener the old configurations are still supported, but you won't get the new features
1113948971 M * Jani The finetuning can come later ;)
1113948988 M * Jani In 6hrs the vserver needs to be up and running, or I have a problem.
1113949023 M * Bertl well, plenty of time left, what do you plan to do until then?
1113949027 M * kalou Jani: lol
1113949039 M * Jani Bertl: Perhaps playing a game ;)
1113949045 M * Jani Or go earlier to bed.
1113949052 M * Doener shouldn't be any problem... remember to adjust the barrier ;) (information on migration are on the alpha util-vserver page)
1113949098 M * Jani *nods* And the procfs :)
1113949105 M * Doener you may also want to modify your <vserver>.sh files to get interface hiding (which is turned off by default with 2.6)
1113949111 M * Doener right!
1113949186 M * Doener Bertl: FYI I'm doing some code auditioning on 2.0pre1
1113949220 M * Jani Btw. the old vserver runs around 140 days without noticing and any complains :)
1113949226 M * Doener is there anything you think I should have a really close look at?
1113949269 M * Bertl Doener: excellent idea ... well .. let me think for a moment ...
1113949295 M * Bertl Jani: so you are already a satisfied customer ;)
1113949328 M * Jani Bertl: Yepe. You would not lose me, when it's not work until the deadline I go back to the old kernel and wait with the kernel patching.
1113949389 M * Bertl do you still have the old tools?
1113949397 M * Jani Yepe.
1113949405 M * Jani *hmms*
1113949405 M * Bertl installed?
1113949410 M * Jani No
1113949415 M * Jani In an backup image
1113949429 M * Bertl okay, you remember the version?
1113949502 M * Jani *hmms* Good question. I think around 0.30-9 or so?
1113949511 M * Jani I think it was the standard Debian vserver
1113949583 M * Jani So. testme.sh is working well now :)
1113949588 M * Jani Thanks much for the help :)
1113949596 M * Doener you're welcome!
1113949619 M * Bertl could you test the 0.12 version for me too?
1113949627 M * Bertl http://vserver.13thfloor.at/Stuff/SCRIPT/testme.sh-0.12
1113949633 M * Jani Sure
1113949634 M * Jani One moment
1113949672 M * Jani [V0.11]  --> I think it should be 0.12 ;)
1113949694 M * Jani *hmms* Same output
1113949696 M * Bertl ahh, I knew I forgot something ;)
1113949712 M * Jani gege
1113949713 M * Jani hehe
1113949734 M * Doener Bertl: ah great, you no longer use a xid that is already used in my setup :)
1113949746 M * Bertl yeah yeah :P ;)
1113950016 M * Jani *hmms* The setattr for the procfs do I need after every reboot of the mainserver?
1113950040 M * Doener Jani: there comes an init-script called vprocunhide with the alpha tools
1113950046 M * Doener just add that to your runlevels
1113950103 M * Jani *nods*
1113950223 M * Jani *jumps up and down*
1113950279 M * Jani It's works :)
1113950285 M * Jani So. Customer can be happy now
1113950289 M * Jani I can more as happy now
1113950297 M * Jani Webservices are up :)
1113950319 M * Doener great!
1113950323 A * Jani gives everyone a beer or cola.
1113950395 M * Doener thanks :)
1113950449 M * Jani *hmms* The new vserver build looks neat :)
1113950497 M * Doener it is, it is
1113950513 M * Jani So. Now I hope not to touch the system for another long time ;)
1113950605 M * Jani Thanks alot for the good work at vserver :)
1113950615 M * Jani It have safe me alot time and trouble :)
1113950739 M * Bertl are you already on our happy vserver users page?
1113950766 M * Bertl (if not, that would be the perfect moment to add yourself ;)
1113950802 M * Jani Oh, I haven't know they exist such page :)
1113950831 M * Bertl http://linux-vserver.org/ (Public Relations)
1113951123 M * Jani So :)
1113951186 M * yarihm n8 everyone
1113951196 M * Bertl night yarihm!
1113951202 M * yarihm night bertl
1113951204 Q * yarihm Quit: Leaving
1113951204 M * kalou good night Yarihm
1113951244 M * Jani Good Night
1113951427 M * kalou Bertl, the call to chbind seems fine, it includes the --cap twice
1113951489 M * Bertl yep, and more important the --secure is specified quite fine
1113951506 M * Bertl (this should remove all dangerous caps)
1113951521 M * kalou so the question is laying in the kernel code ?
1113951521 M * Bertl so .. I have two questions here:
1113951540 M * Bertl a) are you absolutely sure that the other vserver have better flags?
1113951566 M * Bertl b) how did you test the vserver flags (in detail)
1113951593 M * kalou here are the caps for a similar vserver, with exactly the same S_CAPS variable:
1113951594 M * kalou CapInh: 0000000000000000
1113951594 M * kalou CapPrm: 00000000d40c64ff
1113951594 M * kalou CapEff: 00000000d40c64ff
1113951618 M * Bertl okay, that looks fine, now for b)
1113951632 M * kalou to test the vserver CAPs, vserver xxx enter,
1113951642 M * kalou then grep ^Cap /proc/self/status
1113951667 M * Bertl hmm, could you do that with ssh too?
1113951694 M * kalou let me see, I should have a shell there, indeed
1113951796 M * kalou hmm. ok it's right with ssh
1113951869 M * Bertl so I conclude that the 'enter' is somehow wrong there ...
1113951884 M * Bertl do you use static or dynamic contexts?
1113951900 M * kalou hmm. sorry I don't understand the question ;)
1113951927 M * kalou static, I presume
1113951957 M * Bertl okay, what context id does vserver-stat show?
1113951984 M * kalou 49173
1113951992 M * Bertl ah, dynamic contexts ;)
1113952010 M * Bertl well, I'd say, the tools are just broken there ...
1113952028 M * Bertl what is the precise version number and origin?
1113952063 M * kalou  212911 jui  3  2004 util-vserver-0.30.tar.bz2
1113952069 M * kalou c88994e7c22eca1a6a4b14aadca54738  util-vserver-0.30.tar.bz2
1113952103 M * kalou downloaded from Enrico's page
1113952154 M * Bertl hmm, yeah ... I'd file a bug report, if you know the tools have not been modified ...
1113952174 M * kalou did not do any modification, indeed
1113952189 M * kalou thank's for your time Bertl :)
1113952192 M * Bertl but aside from that, the use of dynamic contexts for guests is depreciated
1113952204 M * Bertl kalou: you're welcome!
1113952266 M * kalou In fact, I don't know what distinguishes static contextes or dynamic ones
1113952280 M * kalou Should I use the latest alpha tools instead ?
1113952323 M * Bertl the static ones are static, i.e. they are the same at every vserver start ;)
1113952342 M * Bertl the dynamic ones, OTOH change often ....
1113952404 M * kalou Ok .. the only difference is the way xid get allocated, then ?
1113952429 M * Bertl yes ...
1113952460 Q * mugwump Ping timeout: 480 seconds
1113952495 M * kalou oki
1113952496 M * Doener kalou: are you using fakeinit?
1113952516 M * kalou yep
1113952531 M * Doener and you test that via "vserver xxx enter"?
1113952538 M * kalou only in the incriminated vserver
1113952568 M * Doener "vserver xxx enter" (and probably exec) is broken for fakeinit
1113952576 M * Doener thus you get the full caps
1113952583 M * kalou First test was done this way, indeed. Next Bertl told me to try with ssh, and works ok this way (right caps)
1113952655 M * kalou YES ! that's *exactly* what happens. removing the "fakeinit" flag fixes the problem
1113952669 Q * lilo Read error: Operation timed out
1113952758 M * kalou thank you Doener
1113952766 M * Doener you're welcome
1113952814 M * kalou Ok I'll cope with this !
1113952879 M * Bertl Doener: hmm, maybe we should start a Bug archive?
1113952927 M * kalou btw ... is there any solution to use an advanced scheduling mecanism with 2.4 ?
1113952927 M * Jani So, I go to bed now. After everything is working well here :)
1113952928 M * Doener sounds reasonable
1113952936 M * kalou 'night Jani
1113952937 M * Doener good night Jani!
1113952949 M * Bertl night Jani!
1113952956 M * Jani Good Night everyone :)
1113952959 Q * Jani Quit: Verlassend
1113953215 M * kalou http://linux-vserver.org/Scheduler+Parameters ...
1113953401 M * Bertl kalou: nope
1113953416 M * kalou that's for 2.6 I think
1113953418 M * kalou # sched: Merge scheduler priority of all processes in the vserver
1113953418 M * kalou #        so that it acts a like a single one.
1113953435 M * kalou but this one is for 2.4, right ?
1113953445 M * Bertl yep, that is already done
1113953474 M * Bertl they will be accounted as one process when the 'sched' flag is issued
1113953475 M * kalou I tried to play with "sched" once .. but did not understand the usefulness
1113953505 M * Bertl simple, consider 2 cpu hogs running in vserver A and 20 in vserver B
1113953523 M * kalou yes
1113953548 M * Bertl usually vserver B will get roughly 11 times the amount of cpu resources of vserver A
1113953586 M * Bertl if you issue 'sched' for vserver B, then B will get a little more than 50% of A ;)
1113953635 M * kalou ok. Now it's clear
1113953671 M * Bertl but that doesn't make anything binding ...
1113953713 M * Bertl i.e. it just changes how the scheduler sees the tasks/contexts
1113953769 M * kalou ok. does not wire a new scheduler on the vserver :)
1113953836 M * Bertl precisely, and it has no way to stop any process from scheduling
1113953919 M * kalou ok ... I have two vserver here .. A with 3 cpu hog (dd if=/dev/zero of=/dev/null), B with 1 cpu hog, both "sched" option set
1113953965 M * kalou one dd is taking 95%, and the 3 left are taking around 1.1 each .. :-(
1113953993 M * Bertl dd is no cpu hog ;)
1113954012 M * kalou isn't it ?
1113954024 M * Bertl http://vserver.13thfloor.at/Experimental/TOOLS/cpuhog.c
1113954041 M * Bertl dd is very I/O bound
1113954046 M * kalou ok
1113954103 M * kalou Indeed, I can see much system% use with dd
1113954324 M * kalou Hmmm. A: (sched, 3 cpuhog), B: (sched, 1 cpuhog), the 4 cpuhog are taking around 25%
1113954445 M * Bertl in total? each one?
1113954453 M * kalou each one
1113954504 M * Bertl hmm, unusual .. but I have to admit that I never tested it ...
1113954511 M * kalou :)
1113954526 M * Bertl but nothing I ca not do right now .. hold on
1113954566 M * kalou ok some action great !
1113955003 M * Bertl hmm, yeah, ideed, looks a little strange here ...
1113955011 M * Bertl   PID USER     PRI  NI  SIZE  RSS SHARE STAT %CPU %MEM   TIME COMMAND
1113955011 M * Bertl    14 root      20   0   244  244   200 R    25.0  0.4   0:44 cpuhog
1113955011 M * Bertl    19 root      20   0   244  244   200 R    25.0  0.4   0:23 cpuhog
1113955011 M * Bertl    18 root      19   0   244  244   200 R    24.8  0.4   0:24 cpuhog
1113955011 M * Bertl    20 root      14   0   244  244   200 R    24.4  0.4   0:24 cpuhog
1113955040 Q * Jdogg Ping timeout: 480 seconds
1113955042 M * Bertl $ chcontext --flag sched --xid 100 /tmp/cpuhog &
1113955070 M * Bertl chcontext --flag sched --xid 200 bash -c "/tmp/cpuhog & /tmp/cpuhog & /tmp/cpuhog"
1113955103 A * Bertl is now looking at the code ...