1242174039 J * ard ~ard@shell2.kwaak.net
1242174720 M * Bertl off to bed now ... have a good one everyone!
1242174725 N * Bertl Bertl_zZ
1242178708 Q * pmenier_off Read error: Connection reset by peer
1242178747 J * pmenier_off ~pme@LNeuilly-152-22-8-5.w193-251.abo.wanadoo.fr
1242183943 J * imcsk8 ~ichavero@189.155.63.73
1242188202 Q * imcsk8 Quit: This computer has gone to sleep
1242188284 Q * Sebboh Read error: Connection reset by peer
1242188289 J * Sebboh ~hobbes@ip68-13-125-163.om.om.cox.net
1242189066 Q * Piet Quit: Piet
1242190075 Q * geb Quit: Quitte
1242191498 J * imcsk8 ~ichavero@189.155.63.73
1242191520 Q * imcsk8 
1242193918 Q * AndrewLee Write error: connection closed
1242195995 J * davidkarban ~david@193.85.217.71
1242197780 N * Bertl_zZ Bertl
1242197787 M * Bertl morning folks!
1242198386 J * harobed ~harobed@pda57-1-82-231-115-1.fbx.proxad.net
1242198544 J * doener ~doener@i59F5B560.versanet.de
1242198979 Q * ghislainocfs21 Quit: Leaving.
1242198990 J * ghislainocfs2 ~Ghislain@adsl2.aqueos.com
1242199842 J * cga ~weechat@194.244.1.35
1242200973 J * virtuoso ~s0t0na@81.9.35.135
1242201281 Q * Sebboh 
1242201538 M * pmjdebruijn Bertl: I just tried to apply vs2.3.0.36.4 to linux-2.6.27.23, which works with fuzz=4
1242201547 M * pmjdebruijn otherwise a syscall.32.S file rejects
1242201680 P * virtuoso 
1242203075 J * litw ~litw@123.127.250.67
1242205119 J * gnuk ~F404ror@217.19.56.132
1242206751 J * docelic ~docelic@78.134.194.113
1242207319 J * _nono_ ~gomes@libation.ircam.fr
1242208314 J * coda ~coda@dhcp240-33.satec.es
1242208318 M * coda hi there!
1242208367 M * coda excuse me guys, does anybody know where could I start tunning vserver?
1242208411 M * Bertl what's 'tunning'?
1242208448 M * coda I meant 'tuning' :P
1242208463 M * Bertl okay, how would you like to tune it?
1242208545 M * coda I have an Intel (4xCPU) 3.2ghz and 6gb RAM, and actually I'm hosting 48 linux VPSs there.. actually I needed to upgrade vserver to the latest version and I'm not pretty sure where to set the CPU/Memory limits
1242208582 M * Bertl I presume you are running 64bit on that machine, yes?
1242208596 M * coda yes
1242208621 M * Bertl good, that's already an apropriate tuning
1242208653 M * Bertl with recent kernels (2.6.28/29) you probably want to use the CFS, unless you need hard limits
1242208676 M * Bertl on the memory side, I would just limit the guests to something slightly above the maximum usage so far
1242208718 M * Bertl 48 guests isn't that much, unless each of them is running a java vm :)
1242208796 M * coda hehheheh :)
1242208876 M * coda actually I've started trying to limit them to 128000 blocks (512mb?) in rlimits/as, rlimit/memlock, and rlimit/rss
1242208882 M * coda is that correct?
1242208933 M * Bertl pages, and yes, 4k page size is normal for x86_64
1242208956 M * Bertl note that you do not want to limir as/vm that heavily
1242208961 M * coda oh, sorry, I meant pages.. :P
1242208962 M * Bertl *limit
1242208998 M * Bertl check with /proc/virtual/<xid>/limits for a currently running guest, you'll se what it uses (and used up at max)
1242209031 M * coda yes, actually I've started having problems with my VPSs.. most of them are running out of memory :S
1242209044 M * cehteh memory limiting is some voodoo
1242209100 Q * litw Quit: 暂离
1242209113 M * coda :P
1242209135 M * Bertl it's not that hard, first, consider VM/AS >> RSS >> LOCK
1242209160 M * cehteh some bucket allocator would be nice for vserver
1242209160 M * coda oh no.. sorry about this, probably you're gonna kill me, but I've noticed this server's CPU is not 64bit :(
1242209186 M * Bertl I doubt that, but maybe your distro/kernel is not 64bit
1242209221 M * Bertl what does 'grep flags /proc/cpuinfo' give?
1242209224 M * Bertl (please use paste.linux-vserver.org for everything longer than 3 lines)
1242209254 M * coda flags		: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe pebs bts cid xtpr
1242209325 J * friendly ~friendly@ppp121-44-205-115.lns10.mel4.internode.on.net
1242209415 M * Bertl interesting, and it has 4 cpus and 6GB of memory?
1242209495 M * Bertl cehteh: a bucket allocator for what?
1242209553 M * cehteh memory limits for vservers, based on available memory/priority
1242209591 M * cehteh when no one else uses some memory one vserver could get more
1242209595 M * Bertl well, go ahead, implement it for Linux, I doubt it would make much sense
1242209621 M * cehteh yeah there are plently of problems, see my voodoo quote
1242209660 M * cehteh if you limit memory too much it rather goes into unneccessary swapping .. if you give to much memory .. it starts necessary swapping .. 
1242209663 M * cehteh uhm :)
1242209768 M * cehteh maybe some daemmon which adjust /proc/*/oom_* .. maybe stops processes until memory usage cooled down and so on would be simpler and nicer
1242209832 M * Bertl I don't think that 'stopping' processes will affect the memory usage .... and killing them is what the OOM killer already does
1242209858 M * Bertl and no, on Linux-VServer a memory limit will not result in swapping
1242209871 M * cehteh stopping a process elides it for completely swapped out and then more core memory is available for other things
1242209921 M * cehteh sometimes (often) serializing memory intensive processes in this way provides much better throughput
1242209951 M * cehteh of course if your server is that heavy swapping you may consider something else .. more ram or so
1242209965 M * cehteh but thats not always doable or only transient processes need that much memory
1242210242 J * FireEgl Proteus@WTF.4.1.0.c.0.7.4.0.1.0.0.2.ip6.arpa
1242210251 M * coda sorry, I'm back
1242210288 M * coda Bertl: that's correct: 4 cpus and 6GB of memory
1242210332 M * Bertl well, unfortunate design, a quad core CPU would probably perform twice as fast and use the memory more efficiently
1242210434 M * coda Bertl: ough.. so am I lost? :)
1242210543 M * Bertl well, no, but next time consider this
1242210625 M * coda Bertl: great, thanks :)
1242210786 M * coda Bertl: regarding VM, I'm checking one of my running VPSs and I see it's reaching the limit (45 hits).. so, should I consider increasing that limit?
1242210855 M * coda oh, I have another one that hitted the limit 2128 times :S
1242211192 J * AndrewLee ~andrew@u7.hlc.edu.tw
1242211585 M * coda Bertl: sorry Bertl, just one more quick question.. what could be a good value for the VM value to start trying? should it be a value much higher than 128000 pages?
1242211600 M * coda (sorry for my poor english..)
1242211701 M * Bertl as I said, try without the VM limit, check what the guests use at max (on normal behaviour), and set the limit accordingly
1242211718 M * Bertl the VM varies greatly depending on the processes and their reservations
1242211761 M * coda Bertl: ahhh okok, sorry
1242211811 M * coda Bertl: I really appreciate your help :)
1242211819 J * BenG ~bengreen@94-169-110-10.cable.ubr22.aztw.blueyonder.co.uk
1242211908 M * coda on the other hand, for some reason the swap memory seems to be set to 0.. do you know how to fix that? I mean, it says: total 0, used 0 and free 0
1242211945 M * Bertl the swap inside a guest is the difference between soft and hard RSS limit
1242211971 M * Bertl it has no relation to the swap space on the host, which is maintained completely different
1242212312 M * coda thanks again
1242212501 J * kir ~kir@swsoft-msk-nat.sw.ru
1242213216 Q * docelic Quit: http://www.spinlocksolutions.com/
1242214449 Q * thierryp Remote host closed the connection
1242216253 Q * FireEgl Quit: Leaving...
1242216586 Q * gnuk Ping timeout: 480 seconds
1242216602 J * gnuk ~F404ror@pla93-3-82-240-11-251.fbx.proxad.net
1242217225 M * Bertl off for now ... bbl
1242217229 N * Bertl Bertl_oO
1242218091 Q * friendly Quit: Leaving.
1242218426 Q * harobed Ping timeout: 480 seconds
1242218568 M * harry new test vserver patch online!!! with grsec off course ;)
1242218582 M * harry http://people.linux-vserver.org/~harry/patch-2.6.29.2-vs2.3.0.36.12-grsec2.1.14-20090513.diff
1242218588 M * harry test and report please ;0
1242218589 M * harry ;)
1242219149 M * harry Bertl_oO: is a 2.6.29.3 version coming soon? (there are some "hard" rejects) because of some serious changes in fs_struct...)
1242219793 J * cluk ~cluk@p5B17F878.dip.t-dialin.net
1242220346 J * BenG_ ~bengreen@94-169-110-10.cable.ubr22.aztw.blueyonder.co.uk
1242220349 Q * BenG_ 
1242221492 M * BenG hi all 
1242221526 M * BenG I rolled up an up-to-date kernel for Debian Lenny AMD64:  http://kernels.bristolwireless.net/linux-image-2.6.29.2-vs2.3.0.36.10-beng-amd64_0.1_amd64.deb
1242221534 M * BenG please have a go of this
1242221632 M * BenG I'm having issues I believe are related to having manipulted guests using the 2.6.26 kernel from Debian
1242221686 M * BenG two of my guests will start, but two won't
1242221743 M * BenG vcontext: execvp("/usr/sbin/vspace"): Permission denied
1242221747 M * BenG is the error
1242221775 N * pmenier_off pmenier
1242221813 M * BenG ah, I'm getting that for all the guests now
1242221845 M * BenG vsysctl: open("."): Permission denied
1242221858 M * BenG was the error that stopped only two of them working
1242222301 M * BenG right, I fixed it by fixing the xattrs and then rebooting
1242222308 M * BenG then trying to run my server again
1242222438 M * BenG successful start for 3 of them now, the one that isn't starting hasn't had it's attributes set
1242222474 M * BenG I think I'm talking to myself, but in any case, the rolled kernel works fine, that URL again:
1242222481 M * BenG  http://kernels.bristolwireless.net/linux-image-2.6.29.2-vs2.3.0.36.10-beng-amd64_0.1_amd64.deb
1242222967 J * Piet ~piet@tor-irc.dnsbl.oftc.net
1242223803 J * thierryp ~thierry@home.parmentelat.net
1242224697 Q * BenG Quit: I Leave
1242225041 Q * scientes Ping timeout: 480 seconds
1242225793 J * docelic ~docelic@78.134.201.71
1242225930 J * BenG ~bengreen@94-169-110-10.cable.ubr22.aztw.blueyonder.co.uk
1242225939 Q * BenG 
1242226042 Q * starcode Quit: starcode
1242227062 J * BenG ~bengreen@94-169-110-10.cable.ubr22.aztw.blueyonder.co.uk
1242227623 Q * cluk Quit: Ex-Chat
1242227885 Q * ghislainocfs2 Ping timeout: 480 seconds
1242227937 Q * BenG Quit: I Leave
1242228487 J * ghislainocfs2 ~Ghislain@adsl2.aqueos.com
1242228977 J * BenG ~bengreen@94-169-110-10.cable.ubr22.aztw.blueyonder.co.uk
1242229031 Q * BenG 
1242229616 J * BenG ~bengreen@94-169-110-10.cable.ubr22.aztw.blueyonder.co.uk
1242229665 Q * BenG 
1242229721 Q * cga Quit: got a DELL??? update you BIOS with http://github.com/cga/dellbiosupdate.sh/tree/master ;)
1242229862 J * dowdle ~dowdle@scott.coe.montana.edu
1242230756 Q * balbir_ Ping timeout: 480 seconds
1242231888 Q * kir Quit: Leaving.
1242232045 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d
1242233345 J * Keeper longhorn40@94.50.33.117
1242233957 Q * davidkarban Quit: Ex-Chat
1242234859 J * geb ~geb@earth.gebura.eu.org
1242235076 M * geb hi
1242236375 J * cga ~weechat@82.84.191.5
1242237250 Q * Keeper Quit: Keeper
1242237307 J * doener_ ~doener@i59F5BC39.versanet.de
1242237410 Q * doener Ping timeout: 480 seconds
1242238980 J * balbir_ ~balbir@122.172.37.252
1242239860 Q * gnuk Quit: NoFeature
1242240118 J * SlackLnx ~SlackWare@85.139.11.111
1242240751 Q * larsivi_ Read error: Connection reset by peer
1242240776 J * larsivi ~larsivi@70.84-48-63.nextgentel.com
1242241548 J * hparker ~hparker@2001:470:1f0f:32c:290:96ff:fe50:40fa
1242241796 Q * balbir_ Ping timeout: 480 seconds
1242242471 J * balbir_ ~balbir@122.172.22.28
1242243631 Q * ktwilight Read error: Connection reset by peer
1242243638 J * ktwilight ~ktwilight@247.244-64-87.adsl-dyn.isp.belgacom.be
1242245545 Q * thierryp Quit: ciao folks
1242246302 M * Sebboh- Is there a distro that ships up to date vserver kernels?
1242246321 M * Mr_Smoke Probably not
1242246342 M * Mr_Smoke Well, Gentoo has 2.2.07
1242246348 M * Mr_Smoke That's up-to-date stable.
1242246355 M * Mr_Smoke But 2.3 is still hard masked
1242246357 M * Sebboh- I just realized, why am I using Ubuntu for this? Because the machine had Ubuntu on it before. Screw Ubuntu. :P
1242246369 M * Sebboh- What does hard masked mean?
1242246398 M * Mr_Smoke Means it's in the portage tree but you have to unlock a few security warnings to get it
1242246417 M * Mr_Smoke 2.3 is still considered bleeding edge.
1242246441 M * Mr_Smoke But the latest available is apparently 2.3.0.36.4
1242246468 M * Sebboh- ok.  Never used Gentoo.. Can I build a debian etch (oldstable) vserver inside gentoo with stock vserver 2.2.07?
1242246498 M * Mr_Smoke Most probably, yes
1242246511 M * Mr_Smoke Never done it, but the guest Linux can be pretty much anything I believe
1242246571 J * BenG ~bengreen@94-169-110-10.cable.ubr22.aztw.blueyonder.co.uk
1242246585 Q * BenG 
1242246600 M * Mr_Smoke I need to try that some day
1242246618 M * Mr_Smoke Not that I'm a big fan of Debian (Gentoo all the way !)
1242246689 M * Sebboh- :) Debian rocks. It's.. logical, from the ground up. Well designed.
1242246700 M * Mr_Smoke It's also 4 years old, on average ;)
1242246713 M * Mr_Smoke <tease tease>
1242246722 M * Mr_Smoke Food, brb.
1242246725 M * Sebboh- Well yeah. :) (Why do you think Ubuntu was on that machine!)
1242247066 M * Mr_Smoke Yeah, still need to give Ubuntu Server a shot.
1242247097 J * cga_ ~weechat@82.84.140.253
1242247157 M * zbyniu Sebboh-: pld-linux
1242247175 Q * cga Ping timeout: 480 seconds
1242247256 M * Sebboh- Mr-smoke, no you don't.  Just use Debian at that point..
1242247284 M * Sebboh- zbyniu, ... hey I've heard of that, it came up when I was searching around vserver.
1242247434 M * zbyniu Sebboh-: pld kernels have vserver+grsecurity+apparmor+many netfiler and other patches
1242247436 Q * balbir_ Ping timeout: 480 seconds
1242247525 M * Sebboh- hm. I'm on a single user system. :) Dev box. Just looking for a solution to needing multiple dev boxes.
1242247584 Q * cga_ Ping timeout: 480 seconds
1242247730 M * zbyniu so your problem should be how distro works as guest system, what is on host is not so important
1242247945 M * Sebboh- Hm? Right, I can just use a stock vserver kernel.  Ubuntu doesn't have one, Debian doesn't have one to speak of.. :P I got trapped into the idea of building a vserver kernel for Ubuntu, but yeah, I don't need that to achieve my goal.
1242248018 J * harobed ~harobed@arl57-1-82-231-110-14.fbx.proxad.net
1242248084 M * Mr_Smoke Sebboh-: bah, I was thinking of Ubuntu server as a less-totally-out-of-date Debian in fact
1242248102 M * Mr_Smoke But I was merely speculating. Gentoo makes me very happy so far
1242248120 M * Mr_Smoke And today's fast cpus make it a breeze to use
1242248133 M * Mr_Smoke (compilation time isn't much of a factor anymore)
1242248193 M * Sebboh- Yeah. This is an Athlon (Barton) box so we'll see. :)
1242248258 M * Sebboh- Oh, re: ubuntu server, well..  my experience with Ubuntu over the past week leads me to believe that they rushed Jaunty out, and really focused on the desktop.
1242248290 M * Mr_Smoke Mkay
1242248291 M * Sebboh- The server installation failed repeatedly, I had to manually edit scripts to get grub to install.  But it worked fine for desktop.
1242248299 M * Mr_Smoke Ubuntu sounds like a nice choice for a desktop indeed
1242248309 M * Mr_Smoke I'll be building a media center soon
1242248322 M * Sebboh- oh, yeah, you want mythbuntu for sure. :)
1242248342 Q * bonbons Quit: Leaving
1242249131 M * Mr_Smoke Yeah mythbuntu sounds like a good choice indeed
1242249792 A * hparker likes xbmc for a media box
1242252099 J * ViRUS ~mp@p579B483A.dip.t-dialin.net
1242252575 Q * harobed Ping timeout: 480 seconds
1242252804 Q * SlackLnx Quit: I'll Be Back
1242254865 Q * infowolfe Quit: Leaving
1242255590 Q * geb Ping timeout: 480 seconds
1242255621 Q * dowdle Remote host closed the connection
1242256125 J * geb ~geb@earth.gebura.eu.org
1242256147 J * infowolfe ~infowolfe@c-76-105-242-186.hsd1.or.comcast.net
1242256182 J * scientes ~scientes@97-113-122-26.tukw.qwest.net
1242256825 Q * saulus Read error: Connection reset by peer
1242257266 J * saulus ~saulus@d003031.adsl.hansenet.de