1147219208 M * Radiance i see heh 1147219217 M * derjohn if "mount --bind" a dir to /var/lib/vservers ... where has the barrier flag to be? on the orig dir? or the mount? 1147219254 M * Bertl on /var/lib/vservers/guest/.. (note the ..) 1147219355 M * derjohn yes, obviously per guest. but the ".."? This is the dentry (?) where the flag resides= 1147219370 M * derjohn s/=/?/ 1147219571 M * Bertl hmm? 1147219841 A * derjohn just thought about the ".." you were explicitly mentioning 1147220034 M * Radiance hmm since i set in squid to use 192.168.1.2 it gives this error: Starting proxy server: FATAL: Could not determine fully qualified hostname. Please set 'visible_hostname' 1147220109 M * Radiance setting visible_hostname to all kinda names didn't matter 1147220137 M * derjohn Radiance, # hostname 1147220176 M * derjohn Radiance, 'visible_hostname' is something within squid I think, not the shell 1147220205 M * derjohn Radiance, the hostname of a guest^W vps can be set in /etc/vservers/guests/uts/nodename 1147220212 M * Radiance yeah it doesn't matter what i enter there, the domain, some fake name, it'll keep giving that error 1147220217 M * Radiance ok lemme check 1147220702 M * Radiance derjohn, didn't do anything 1147220721 M * Radiance /etc/hosts or dns might have to do something with this ? 1147220739 M * Radiance it only happened after i told squid to listen on 192.168.1.2 instead of the internet ip 1147220756 M * Bertl check /etc/hosts 1147220769 M * Radiance ok 1147220776 M * Bertl if it doesn't contain an entry for 192.168.1.2, you probably need that 1147220843 M * Radiance the hosts in the guest or host btw ? 1147220868 M * Bertl guest, everything guest specific (like lookups) happen inside the guest 1147220889 M * Radiance ok 1147220969 M * Radiance ok that worked like a charm 1147220980 M * Bertl the mechanics behind the networking (as it is right now) are basically: 1147220982 M * Radiance didn't know that squid depends so much on the /etc/hosts file 1147220997 M * Bertl the first assigned ip is the _primary_ 1147221012 M * Bertl 127.0.0.1 is 'remapped' to that one 1147221029 M * Bertl binding to 0.0.0.0 will bind all assigned ips 1147221036 M * Bertl (that's all) 1147221085 M * Bertl well, squid seems to 'resolve' the ip 1147221113 M * Bertl so, as your dns service probably doesn't reverse 192.168.x.x, you need to specify it there 1147221200 M * Radiance yeah, i've been using squid a long time, but never thought about it this way heh 1147221283 M * Radiance btw, the host has only 1 interface which is eth0, when i assign in /etc/vservers/guest/interfaces a new section, then dev must contain in this situation eth0 unless i want to use for example a dummy device or similar ? 1147221318 M * Bertl yes, or alternatively nodev, if you assign the ip manually 1147221341 M * Radiance ah perfect 1147221354 M * Radiance just want to understand these things the right way hehe 1147221367 M * Radiance that flower page btw ... it almost blinded me here in the night ;) 1147221380 M * Bertl hehe, what css stylesheet? :) 1147221418 M * Radiance yellow foreground, pink background with a special type of plant ;) 1147221431 M * Radiance plus blue and green 1147221433 M * Radiance :-) 1147221602 M * Bertl ah, well, I usually use weedpage ... quite nice 1147221682 M * derjohn sine some rcs I get "hanging shells" in guests sometimes. I can immediately re-logon, but the one shell is lost. 1147221739 M * derjohn ah, mainly when there is excessive IO on the machine 1147221776 M * derjohn What the time ... of .. off now, bye ! 1147221803 M * Radiance laters derjohn :) 1147221875 J * pengyong ~lala@222.188.131.245 1147221898 Q * pengyong Quit: 1147222179 Q * doener_ Quit: leaving 1147222324 Q * michal` Ping timeout: 481 seconds 1147222359 M * Bertl okay, off to bed too .. back tomorrow ... 1147222368 M * Bertl have a good one everyone ... cya! 1147222375 N * Bertl Bertl_zZ 1147222483 M * Skram so.. anyone think http://linuxmanaged.com/ is worth any $$? 1147222588 J * michal` ~michal@www.rsbac.org 1147223648 M * [PUPPETS]Gonzo well, no? 1147223772 M * Skram ? 1147223821 M * hillct may be worth something 1147223831 M * hillct but its' a niche thing 1147223858 M * hillct you might be better off creating a portal to other embedded linux projects and devices 1147223884 M * hillct just put up a directory 1147225110 M * Skram nah 1147225114 M * Skram what a waste. 1147225934 J * ntrs ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com 1147229186 M * Radiance guys, i added that 192.168.1.2 to the guest as a second, i noticed when i do within the guest ifconfig then it shows that ip and not the primary. Also when i try to ssh to the guest from a remote box then the sshd of the host responds but not the sshd of guest. Can some one see what i'm missing here ? 1147229234 M * Radiance an ip addr does show them all as they should 1147229408 M * Skram make sure that the host ssh server only binds to the HOST ip. 1147229455 M * Radiance yup it does 1147229465 M * Radiance binds to the host ip 1147229475 M * Radiance the guest sshd binds to the assigned ip 1147229525 M * Radiance i added 192.168.1.2 as the second ip for the guest, when i do ifconfig it shows this ip and not the primary ip (internet ip) for the guest anymore 1147229547 M * Radiance i'm sure this is what is causing me not to be able to login to the guest anymore and the host sshd responds instead 1147229598 M * hillct Radiance: sounds like the host sshd is still binding to all IPs 1147229616 M * hillct edit /etc/ssh/sshd.conf and replace * with the host IP 1147229620 M * hillct restart the guests 1147229629 M * hillct er restart host sshd 1147229634 M * hillct restart guests 1147229646 M * hillct guest sshd will then bind to guest IP 1147229669 Q * s0undt3ch Ping timeout: 480 seconds 1147229730 M * Radiance hmm not the host sshd, for some reason it escaped me that bind is binded to all ips on the host 1147229735 M * Radiance so this prolly is the problem 1147229756 M * hillct yup 1147229765 M * hillct default is for sshd to bind to all IPs 1147229848 M * hillct likewise, if you run sendmail in a guest, it'll try to bind to 127.0.0.1 which will fail 1147229862 M * hillct simple edit of sendmail.cf deals with it 1147230499 M * Radiance yeah, i got bind to listen on the ip of the host, but the rdnc port and 2 other ports belonging also to bind are still on all ips 1147230862 M * Radiance ok rndc is now listening on loopback 1147230868 M * Radiance don't need it anyway for now 1147231028 M * Radiance ok that solved it 1147231242 M * Radiance thanks, off for some sleep now :) 1147231637 J * f_ ~f_@83-215-237-1.seek.stat.salzburg-online.at 1147231720 Q * f_ Quit: 1147233473 J * s0undt3ch ~s0undt3ch@bl7-240-201.dsl.telepac.pt 1147233815 Q * VAndreas jupiter.oftc.net neutron.oftc.net 1147233815 Q * jkl jupiter.oftc.net neutron.oftc.net 1147233903 J * jkl eric@c-71-56-216-223.hsd1.co.comcast.net 1147233903 J * VAndreas ~Hossa@212.110.98.7 1147233926 Q * softi42 jupiter.oftc.net unununium.oftc.net 1147233926 Q * Zaki jupiter.oftc.net unununium.oftc.net 1147233926 Q * TrueLight jupiter.oftc.net unununium.oftc.net 1147233926 Q * shedi jupiter.oftc.net unununium.oftc.net 1147233926 Q * odedra jupiter.oftc.net unununium.oftc.net 1147233926 Q * nebuchadnezzar jupiter.oftc.net unununium.oftc.net 1147233926 Q * virtuoso jupiter.oftc.net unununium.oftc.net 1147233926 Q * dsoul jupiter.oftc.net unununium.oftc.net 1147233926 Q * brc_ jupiter.oftc.net unununium.oftc.net 1147233926 Q * kir jupiter.oftc.net unununium.oftc.net 1147233926 Q * bubulak jupiter.oftc.net unununium.oftc.net 1147233926 Q * h01ger jupiter.oftc.net unununium.oftc.net 1147233926 Q * jake- jupiter.oftc.net unununium.oftc.net 1147233926 Q * pusling jupiter.oftc.net unununium.oftc.net 1147233926 Q * Greek0 jupiter.oftc.net unununium.oftc.net 1147233926 Q * redtux jupiter.oftc.net unununium.oftc.net 1147233926 Q * mugwump jupiter.oftc.net unununium.oftc.net 1147233926 Q * mcp jupiter.oftc.net unununium.oftc.net 1147233926 Q * Hollow jupiter.oftc.net unununium.oftc.net 1147233926 Q * morrigan jupiter.oftc.net unununium.oftc.net 1147233926 Q * nokoya jupiter.oftc.net unununium.oftc.net 1147234018 J * softi42 ~softi@p549D583C.dip.t-dialin.net 1147234018 J * Zaki ~Zaki@212.118.96.92 1147234018 J * TrueLight ~kvirc@truelight.xs4all.nl 1147234018 J * odedra ~oded@bzq-84-108-241-136.cablep.bezeqint.net 1147234018 J * shedi ~siggi@inferno.lhi.is 1147234018 J * nebuchadnezzar ~nebu@zion.asgardr.info 1147234018 J * virtuoso ~s0t0na@shisha.spb.ru 1147234018 J * dsoul darksoul@vice.ii.uj.edu.pl 1147234018 J * brc_ bruce@20151237032.user.veloxzone.com.br 1147234018 J * kir ~kir@swsoft-mipt-nat.sw.ru 1147234018 J * bubulak ~bubulak@cicka.wnet.sk 1147234018 J * h01ger ~holger@socket.layer-acht.org 1147234018 J * jake- qscand@murlocs.org 1147234018 J * pusling pusling@195.215.29.124 1147234018 J * redtux ~redtux@pc199.pub.univie.ac.at 1147234018 J * Greek0 ~greek0@85.255.145.201 1147234018 J * mugwump ~samv@watts.utsl.gen.nz 1147234018 J * mcp ~hightower@wolk-project.de 1147234018 J * Hollow ~hollow@home.xnull.de 1147234018 J * nokoya young@hi-230-82.tm.net.org.my 1147234018 J * morrigan morrigan@212.16.62.52 1147235130 Q * VAndreas jupiter.oftc.net neutron.oftc.net 1147235130 Q * jkl jupiter.oftc.net neutron.oftc.net 1147235219 J * jkl eric@c-71-56-216-223.hsd1.co.comcast.net 1147235219 J * VAndreas ~Hossa@212.110.98.7 1147237964 J * _coocoon_ ~coocoon@p54A0573E.dip.t-dialin.net 1147237973 M * _coocoon_ morning 1147238063 M * Skram hello 1147238315 J * f_ ~f_@83-215-237-1.seek.stat.salzburg-online.at 1147238947 Q * f_ Quit: This computer has gone to sleep 1147239241 J * f_ ~f_@83-215-237-1.seek.stat.salzburg-online.at 1147239683 J * MrX ~urk@219.95.239.89 1147242219 J * replayss ~pirch@61.7.137.47 1147242222 J * dsoul_ darksoul@vice.ii.uj.edu.pl 1147242233 P * replayss 1147242512 Q * f_ Quit: This computer has gone to sleep 1147242577 Q * dsoul Ping timeout: 480 seconds 1147242720 J * f_ ~f_@83-215-237-1.seek.stat.salzburg-online.at 1147242999 Q * f_ Quit: 1147243262 J * f_ ~f_@83-215-237-2.seek.stat.salzburg-online.at 1147243386 N * dsoul_ dsoul 1147243559 J * chelli ~thomas@PC1.ideolabs.TGZ-Ilmenau.de 1147246154 J * dna ~naucki@dialer-171-162.kielnet.net 1147246449 J * pagano_ ~pagano@lappagano.cnaf.infn.it 1147246482 N * otaku42_away otaku42 1147247987 J * dsoul_ darksoul@vice.ii.uj.edu.pl 1147248404 Q * dsoul Ping timeout: 480 seconds 1147248739 Q * dsoul_ Ping timeout: 480 seconds 1147249044 J * dsoul darksoul@vice.ii.uj.edu.pl 1147250211 Q * MrX Ping timeout: 480 seconds 1147250429 J * LiNuX ~LiNuX@85.104.63.92 1147250498 M * LiNuX hi? 1147250617 T * LiNuX 4Www.chatmerkezi.com 1147250622 T * LiNuX Www.chatmerkezi.com 1147250689 M * [PUPPETS]Gonzo boah junge, geh' kacken... 1147250732 T * [PUPPETS]Gonzo http://linux-vserver.org/ | latest stable 2.01, 1.2.10, 1.2.11-rc1, devel 2.1.0, exp 2.{0.2,1.1}-rc19 | util-vserver-0.30.210 | libvserver-1.0.2 & vserver-utils-1.0.3 | He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the wiki, and we'll forget about the minute ;)' 1147250750 M * LiNuX [PUPPETS]Gonzo!!! 1147250755 M * LiNuX Akill! 1147250852 T * LiNuX http://www.chatmerkezi.com/chat.asp 1147250857 T * [PUPPETS]Gonzo http://linux-vserver.org/ | latest stable 2.01, 1.2.10, 1.2.11-rc1, devel 2.1.0, exp 2.{0.2,1.1}-rc19 | util-vserver-0.30.210 | libvserver-1.0.2 & vserver-utils-1.0.3 | He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the wiki, and we'll forget about the minute ;)' 1147250860 M * [PUPPETS]Gonzo FUCK OFF? 1147250866 T * LiNuX http://www.chatmerkezi.com/chat.asp 1147250870 M * LiNuX akill!! 1147250875 M * LiNuX root Owner. 1147250902 M * LiNuX Administrative info about reticulum.oftc.net-LiNuX LiNuX@Security.system 1147250905 Q * shedi Quit: Leaving 1147250905 M * LiNuX ok? 1147250909 M * LiNuX Link chatmerkezi.. 1147250910 M * LiNuX pls. 1147250928 T * [PUPPETS]Gonzo http://linux-vserver.org/ | latest stable 2.01, 1.2.10, 1.2.11-rc1, devel 2.1.0, exp 2.{0.2,1.1}-rc19 | util-vserver-0.30.210 | libvserver-1.0.2 & vserver-utils-1.0.3 | He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the wiki, and we'll forget about the minute ;)' 1147250935 T * LiNuX http://www.chatmerkezi.com/chat.asp 1147250967 M * dna rofl 1147250970 M * dna mode +t 1147250990 M * [PUPPETS]Gonzo +b would help, too 1147250997 M * LiNuX Link... 1147251001 M * LiNuX Administrative info about reticulum.oftc.net-LiNuX LiNuX@Security.system 1147251006 M * LiNuX http://www.chatmerkezi.com/chat.asp' 1147251015 M * LiNuX CR.2.1.2 1147251051 M * BenBen wtf? 1147251055 M * BenBen hey all 1147251061 M * [PUPPETS]Gonzo hi benben 1147251096 M * [PUPPETS]Gonzo wird Zeit, dass Bertl aufwacht :-| 1147251221 M * BenBen das kann ja noch dauern... 1147251259 T * BenBen http://linux-vserver.org/ | latest stable 2.01, 1.2.10, 1.2.11-rc1, devel 2.1.0, exp 1147251263 M * BenBen 2.{0.2,1.1}-rc19 | util-vserver-0.30.210 | libvserver-1.0.2 & vserver-utils-1.0.3 | He who asks a question is a fool for a minute; he who doesn't 1147251267 T * [PUPPETS]Gonzo http://linux-vserver.org/ | latest stable 2.01, 1.2.10, 1.2.11-rc1, devel 2.1.0, exp 2.{0.2,1.1}-rc19 | util-vserver-0.30.210 | libvserver-1.0.2 & vserver-utils-1.0.3 | He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the wiki, and we'll forget about the minute ;)' 1147251272 M * BenBen thx :) 1147251289 M * BenBen topic-war: the new game for the whole family :) 1147251411 F * ChanServ +o daniel_hozac 1147251454 F * daniel_hozac +b *!*@85.104.63.92 1147251456 K LiNuX daniel_hozac LiNuX 1147251465 F * daniel_hozac -o daniel_hozac 1147252016 M * [PUPPETS]Gonzo thx 1147252062 J * insomniac ~insomniac@slackware.it 1147252064 M * insomniac hi 1147252124 J * Milf ~Miranda@ipsio25.ipsi.fraunhofer.de 1147252162 M * daniel_hozac hello 1147253171 J * MrX ~urk@219.95.26.32 1147253606 J * shedi ~siggi@tolvudeild-201.lhi.is 1147253996 M * derjohn since which rc do we have that capability 'masking' ? 1147254024 M * daniel_hozac -rc19 1147254042 M * daniel_hozac umm, -rc18. 1147254054 M * derjohn daniel_hozac, countdown runs ;) 1147254104 M * derjohn in devel only? or ported to stable rc , too? 1147254112 M * daniel_hozac devel only, for now. 1147254630 T * * http://linux-vserver.org/ | latest stable 2.01, 1.2.10, 1.2.11-rc1, devel 2.1.0, exp 2.{0.2,1.1}-rc19 | util-vserver-0.30.210 | libvserver-1.0.2 & vserver-utils-1.0.3 | He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the wiki, and we'll forget about the minute ;)' 1147254630 T * [PUPPETS]Gonzo - 1147254637 Q * tokkee Remote host closed the connection 1147254643 M * daniel_hozac (because they were given to lusers in the guests too) 1147254649 J * tokkee tokkee@ssh.faui2k3.org 1147254686 M * derjohn well if I do a 'ip addr add' ... the context will have the ccapability (?) to add the ip , there wont be an error, but the ip does not come up? 1147254702 M * daniel_hozac there will be an error, because the context lacks the capability. 1147254714 M * daniel_hozac the process has the capability though. 1147254784 M * derjohn why doesnt starting bind in a context give an error? because bind actually doesnt use it's 'capa' ? 1147254796 M * daniel_hozac exactly. 1147254816 M * derjohn and ccapability is "context" capability? 1147254832 M * daniel_hozac ccapabilities is Context Caps on http://linux-vserver.org/Caps+and+Flags 1147254839 Q * f_ Quit: This computer has gone to sleep 1147254845 M * daniel_hozac bcapabilities are the regular capabilities, but that the context possess. 1147254927 M * daniel_hozac (using context capabilities earlier was dumb and confusing. i should've said context's bcaps all along) 1147254956 J * f_ ~f_@83-215-237-1.seek.stat.salzburg-online.at 1147255329 Q * hillct Ping timeout: 480 seconds 1147255739 M * derjohn ccaps are those a process within a context posesses, bcaps are the caps a context posesses? *confusing* , but ordered form top to buttom smowhow 1147256759 M * pagano_ to change runtime the cpu freq of vm ? 1147256807 M * pagano_ I remember in bert's video tutorial that he chenged a file... 1147256817 M * daniel_hozac derjohn: no. 1147256825 M * daniel_hozac ccaps and bcaps are both context properties. 1147256860 M * daniel_hozac ccaps are special "slices" of larger capabilities, like allowing secure mounts, ping, etc. 1147256870 M * derjohn **confused** 1147256871 M * daniel_hozac bcaps are regular capabilities, but for the context. 1147256880 M * derjohn aaaahhhhhhh 1147256917 M * derjohn b stand for ... ? the letter after 'a' ? 1147256929 M * derjohn sry .. character after a? 1147256930 M * daniel_hozac i have no idea... 1147256944 M * derjohn *lol* bertl's capas ... 1147256982 M * daniel_hozac haha, perhaps. 1147257002 M * daniel_hozac pagano_: vsched. 1147257101 M * pagano_ thank u 1147257137 M * derjohn softi42, bcapas are normal capas assigned to a whole context, while ccapas are more fine granular 'rights'. to really use the ccapas you have to give the bcapa to the conext. 1147257159 M * derjohn sry, softi42, I always want to type "so" and xchat is completing your name ;) 1147257161 M * daniel_hozac no. 1147257183 M * daniel_hozac the purpose of the ccaps is that you don't have to give the whole bcap to the context. 1147257196 M * daniel_hozac the process needs to have the capability though. 1147257202 M * derjohn hm? I thought I cannot the confiure ccapas at all? 1147257203 M * pagano_ yes http://linux-vserver.org/CPU+Scheduling is empty 1147257245 M * daniel_hozac umm, ccapabilities is in the configuration... 1147257253 M * derjohn pagano_, http://linux-vserver.org/Scheduler+Parameters ? 1147257280 M * daniel_hozac http://linux-vserver.org/vsched+explained 1147257280 A * derjohn re-reads the wiki .... 1147257325 M * pagano_ oh well good point to start 1147258342 M * Hollow morning 1147258365 M * daniel_hozac morning. 1147258554 N * VooDooMaster[A] VooDooMaster 1147258554 A * VooDooMaster is no longer away : Gone for 16 hours 43 minutes 18 seconds 1147258571 J * doener ~doener@i5387E4C5.versanet.de 1147258634 M * daniel_hozac morning doener! 1147258689 M * doener morning :) 1147259034 M * pagano_ uff, to see if it works: 1147259056 M * pagano_ a --> cat /proc/virtual/49186/sched 1147259088 M * pagano_ with the new values that I have decided is enough 1147259091 M * pagano_ ? 1147259113 M * pagano_ here I can see the new FillRate, interval etc 1147259120 M * daniel_hozac yeah, but you'll need sched_prio/sched_hard too. 1147259161 M * doener hm, isn't sched_prio set by default? 1147259164 M * pagano_ [root@vm1 ~]# cat /etc/vservers/vm5/flags 1147259164 M * pagano_ sched_prio 1147259174 M * doener or did that change? 1147259190 M * pagano_ yes, a file called "flags" with inside one line: sched_prio ? 1147259217 M * daniel_hozac not set by default here, at least. 1147259222 M * daniel_hozac right. 1147259223 M * derjohn pagano_, yes, you have tp set that manually 1147259224 M * doener ok 1147259293 M * pagano_ uhmm it seems to me that I don't have different performance 1147259346 M * pagano_ I'm doing some tests. 1147259747 M * pagano_ can I refresh these values without restarting the vm? 1147259770 M * doener yes 1147259846 M * pagano_ how ? :-) 1147259974 M * doener using vsched, see command line help for details 1147260016 M * pagano_ ah k 1147260114 J * doener_ ~doener@i5387E225.versanet.de 1147260119 J * dsoul_ darksoul@vice.ii.uj.edu.pl 1147260445 M * pagano_ I have 2 cpu in hyper thread so 4 kernel cpu 1147260474 M * pagano_ I want a very low cpu usage for a vm 1147260505 M * pagano_ I have tried: fill rate 1 and fill interval 20 1147260518 M * pagano_ but the vm load is still high 1147260528 Q * doener Ping timeout: 480 seconds 1147260531 Q * dsoul Ping timeout: 480 seconds 1147260551 M * pagano_ while true ; do echo "caz" >/dev/null; done 1147260562 M * pagano_ this is my hot_cpu test :) 1147260634 M * doener_ sched_prio only adjust the priority of processes in the vserver. if you want to see sth. like "using only 25% cpu" you need sched_hard 1147260656 M * doener_ and in recent versions even with sched_hard a vserver may use 100% if the others are idle 1147260765 M * pagano_ I have understood, I use last version / last kernel 1147260797 M * pagano_ so just a minute for the response 1147260998 N * dsoul_ dsoul 1147261064 M * pagano_ now It works because a top show me low values 1147261085 M * pagano_ but ofter the vm stucks 1147261096 M * pagano_ often 1147261201 N * Bertl_zZ Bertl 1147261206 M * Bertl morning folks! 1147261246 M * doener_ morning Bertl 1147261273 Q * Johan Ping timeout: 480 seconds 1147261285 J * Oli_ ~skycode@212.224.224.150 1147261291 M * Bertl welcome Oli_! 1147261299 M * Oli_ hi folks 1147261346 M * pagano_ hi Bertl ;) 1147261350 M * Bertl doener_: so if I got that right, 2.6.16.15 is still sensitive to the lock issue, yes? 1147261407 M * Bertl hey pagano_! btw, your while loop is not very good for testing 'cpu' load, you might consider something more cpu bound, but it is probably fine for testing 'default' load ... 1147261418 J * hehehe- hehe@62.150.166.236 1147261424 M * Bertl welcome hehehe-! 1147261462 M * pagano_ bert send me yout hot_cpu script :p 1147261556 M * daniel_hozac http://vserver.13thfloor.at/Experimental/TOOLS/cpuhog.c 1147261564 M * daniel_hozac Bertl: correct. 1147261570 M * pagano_ ohh great :) 1147261840 M * Bertl hehehe-: what's up? 1147262091 J * hallyn ~xa@adsl-75-0-155-101.dsl.chcgil.sbcglobal.net 1147262555 M * harry daniel_hozac: what's wrong with: int main(){for(;;);} ???? 1147262595 M * Bertl welcome hallyn! 1147262694 Q * hehehe- Read error: Connection reset by peer 1147262806 M * daniel_hozac harry: it's not my code :) 1147262864 J * Johan ~finger@lounge.datux.nl 1147262869 M * Bertl wb Johan! 1147262899 M * hallyn morning 1147262923 M * hallyn Bertl: fyi, I invoked your Name in an email on linux-kernel this morning 1147262945 M * Bertl yep, just read it :) 1147262966 M * Bertl will comment on that to explain some cases ... 1147262992 M * phreak`` morning doener_, Bertl, hallyn, harry, daniel_hozac .. 1147263009 M * derjohn m0in, phreak`` 1147263011 M * daniel_hozac morning phreak`` 1147263025 M * Bertl hey phreak``! 1147263051 M * hallyn morning 1147263453 Q * chelli Remote host closed the connection 1147263490 J * id23 ~id@p54A01DAF.dip0.t-ipconnect.de 1147263495 M * id23 hi Bertl 1147263508 M * id23 hi #vserver 1147263550 M * daniel_hozac hello 1147263676 J * gypsymauro ~Io@host103-11.pool21757.interbusiness.it 1147263679 M * gypsymauro hello 1147263701 M * gypsymauro there is a way on a guest to set a route that isn't available in the root? 1147263702 M * Bertl welcome gypsymauro! 1147263736 M * Bertl gypsymauro: yes and no, if you are looking for a separate gateway for your guest, that can be done (easily) 1147263762 M * Bertl gypsymauro: but as networking happens on the host, the host can _always_ use a configured route 1147263802 M * gypsymauro Bertl: uhm I've to add a specific route on the guest that will not be available on the root 1147263817 M * gypsymauro or.. 1147263831 M * gypsymauro just root and this specific guest can use:) 1147263878 M * Bertl that is easily done by utilizing routing tables 1147263895 M * gypsymauro there is an example somewhere? 1147263915 M * Bertl just put your specific route into a separate table, and set a rule to use it for your guest's IPs 1147263923 M * Bertl yes, there is an example, sec 1147263924 M * daniel_hozac http://archives.linux-vserver.org/200311/0470.html 1147264092 M * Bertl btw, should we make a wiki page for articles and such? 1147264103 M * Bertl just stumbled over this for example: http://ieeexplore.ieee.org/xpl/freeabs_all.jsp?arnumber=1430092 1147264137 M * daniel_hozac sure. 1147264182 M * daniel_hozac you wouldn't happen to be an IEEE member? :) 1147264324 J * harti ~hw@83-215-237-5.seek.stat.salzburg-online.at 1147264437 M * Bertl daniel_hozac: well, that would not allow to put the article online AFAIK 1147264487 M * Bertl btw, members pay $13 while non-members have to pay $35 1147264507 M * daniel_hozac oh. 1147264567 M * Bertl it's a weird thing ... 1147264712 Q * id23 Remote host closed the connection 1147264767 M * Bertl wb harti! 1147264803 M * harti hello 1147265114 Q * Oli_ Quit: Oli_ 1147265821 Q * _coocoon_ Quit: KVIrc 3.2.0 'Realia' 1147266218 M * harry anyone here? 1147266327 M * daniel_hozac probably. 1147266370 M * daniel_hozac Bertl: do you know what fs/binfmt_elf_fdpic.c is for? 1147266429 M * Bertl not yet 1147266470 M * daniel_hozac ah, FRV specific. 1147266545 M * daniel_hozac am i missing something, or is line 469 too late? 1147266628 Q * VooDooMaster Quit: Do fish get thirsty? 1147266722 M * harry kinky 1147266732 M * harry can i install a redhat AS 3 in a vserver? 1147266735 M * Bertl daniel_hozac: should work, but it's weird code 1147266754 M * daniel_hozac Bertl: hmm? how could line 465 work? 1147266759 M * Bertl harry: why not? 1147266791 M * harry Bertl: how would i do that? 1147266806 M * Bertl daniel_hozac: hmm, good point! 1147266822 M * Bertl harry: well, doesn't it use RPMs? 1147266843 M * harry yes, it does 1147266844 M * harry :) 1147266873 M * daniel_hozac harry: so... -m rpm? 1147266910 M * daniel_hozac or you could just create an apt/yum repo and use that. 1147266922 A * harry still thinks there should be a clearer layout for the vserver page.. :S 1147266951 M * daniel_hozac feel free ;) 1147266959 M * harry i can't do that 1147266969 M * harry people that are into development should do that 1147266991 M * harry there a re a LOT of old pages/links on the main page 1147266995 M * daniel_hozac document and create layout? 1147266999 M * harry i can't find links to grsec 1147267024 M * harry it's there somewhere, but as allways: i don't know how to reach important pages (as examples on how it works 1147267078 M * Bertl harry: google is your friend 1147267089 M * daniel_hozac searching for grsec on the front page takes you to Tools and patches, where you've linked your patches. 1147267093 M * harry yes, but a clear layout on the vserver page would be nice too 1147267173 M * harry http://linux-vserver.org/grsecurityHowto 1147267176 M * harry totally outdated btw 1147267192 M * Bertl well, that would not change with a different layout, would it? 1147267204 M * harry no, but a cleanup would be nice 1147267215 M * harry cleanup + redesign of layout 1147267222 M * harry but as i said... i can't do that 1147267228 M * harry i don't know enough about vserver 1147267236 M * harry what's useful etc.. 1147267244 M * Hollow learn by doing :) 1147267317 M * daniel_hozac why would you have to know about vserver? 1147267327 M * daniel_hozac if you're just organizing it... 1147267356 M * harry yes, but i don't necessarily know who still uses 2.4 stuff e.g. 1147267362 M * harry or 1.x.x vserver 1147267369 M * Bertl eyck :) 1147267390 M * doener_ and only eyck(?) ;) 1147267396 M * Bertl (just kidding :) 1147267397 M * harry if i find the time, i'll see if i can change it... i'll make a linux-vserver/harry or so ;) 1147267410 M * harry which, someday, might be changed then ;) 1147268031 M * Bertl daniel_hozac: do we have an all-in-one patch for the lock issues with/for Trond's patches? 1147268355 M * daniel_hozac http://daniel.hozac.com/vserver/delta-locks-fix02.diff includes Trond's fix. 1147268626 M * Bertl ah, good! 1147268740 M * Bertl do we have any other remaining issues for *rc19 1147268849 M * daniel_hozac http://daniel.hozac.com/vserver/delta-utsname-fix01s.diff 1147268917 M * Bertl ah, cool, btw, unwanted .orig :) 1147268934 M * daniel_hozac whoops. 1147268977 M * daniel_hozac i'm still running find-problems on the devel tree, to make sure i didn't forget any includes :) 1147269307 M * Bertl btw, it looks like it might be worth breaking out the vx_new_utsname() and move that into a separate include file (and maybe include it with the definition of system_utsname? 1147269365 M * daniel_hozac well, vx_new_utsname is the reason most things include vs_cvirt. 1147269381 M * Bertl that's what I mean 1147269387 M * daniel_hozac http://daniel.hozac.com/vserver/delta-utsname-fix01d.diff 1147269463 M * daniel_hozac i guess including it from linux/utsname.h wouldn't hurt. 1147269526 M * Bertl yes, but we don't want to make a new dependancy there for the other stuff (which is not uts related) 1147269539 M * Bertl granted that uts is one of the main purposes 1147269617 M * daniel_hozac does (de)activate_task/uninterruptible_(inc|dec) need to be in a header at all? 1147269740 M * daniel_hozac it seems they're only used in kernel/sched.c 1147269778 M * Bertl probably not, we might be able to move it into sched_hard or similar 1147269919 M * daniel_hozac so if we move those, we could just drop the includes and rename it to vs_utsname. 1147269991 M * daniel_hozac (and then include it from utsname.h) 1147270565 M * Bertl sounds like a plan! 1147270985 Q * harti Quit: Leaving 1147271493 Q * f_ Quit: This computer has gone to sleep 1147271697 Q * pagano_ Ping timeout: 480 seconds 1147272011 Q * mire Ping timeout: 480 seconds 1147272567 J * mire ~mire@120-167-222-85.COOL.ADSL.VLine.Verat.NET 1147272753 P * gypsymauro 1147272755 J * yarihm ~yarihm@80-218-6-5.dclient.hispeed.ch 1147272761 M * daniel_hozac Bertl: hmm, including it from utsname doesn't work. 1147272820 M * daniel_hozac sched includes vs_base.h, which includes context.h, which includes cvirt_def.h, which includes utsname.h, which leads to incomplete types as current is dereferenced. 1147272888 M * Loki|muh is it a common problem that lvm volumes are unremovable without reboot when they contained a vserver-guest? 1147272904 Q * mkhl Ping timeout: 480 seconds 1147272957 M * Loki|muh -comon+ 1147272960 M * Loki|muh known 1147272967 M * daniel_hozac was it mounted when another guest was started? in which namespace are you mounting it? 1147273073 M * Loki|muh the lvm are mounted at system start with the host-namespace 1147273092 M * Loki|muh each vserver has a special volume 1147273101 M * daniel_hozac and do you still have any guests running? 1147273106 M * Loki|muh yes 1147273112 M * daniel_hozac that's why then. 1147273124 M * Loki|muh but they have their own volumes 1147273146 M * daniel_hozac it doesn't matter, as their namespace is created, all the host's mounts are copied. 1147273151 M * Bertl daniel_hozac: will look into it, might require some massaging 1147273172 M * Loki|muh uh - oh... :( 1147273185 M * Loki|muh is there are workaround or anything I can do? 1147273197 M * Bertl just enter the namespaces (vnamespace) and unmount it 1147273232 M * Loki|muh foreach ( namespace ) { vnamespace umount volume }? 1147273278 M * daniel_hozac pretty much. 1147273759 M * daniel_hozac waldi: re namespace-cleanup: is it really supposed to be after the mounting? 1147273796 M * daniel_hozac waldi: sorry, i'm stupid. 1147273852 M * Bertl hmm? 1147273887 M * daniel_hozac http://daniel.hozac.com/vserver/util-vserver/util-vserver-0.30.210-namespace-cleanup.patch 1147273943 M * derjohn daniel_hozac, was a mistake contacting w-a-l-d-i ? it's very unlikely that he answers at all ... 1147273966 M * daniel_hozac hmm? how so? 1147273972 M * Bertl derjohn, daniel_hozac: hmm, what am I missing? 1147273996 M * derjohn daniel_hozac, just my experience .... 1147274023 M * derjohn daniel_hozac, I tried to contact him here and on #debian-kernel ... 1147274045 M * daniel_hozac well, i've talked to him on here before about his patches. 1147274065 M * derjohn daniel_hozac, always with the same result. So: I started to build my own kernel debs again. 1147274082 M * derjohn daniel_hozac, that must be long ago? 1147274369 Q * Milf Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org 1147274384 M * daniel_hozac derjohn: beginning of april. 1147274533 M * derjohn daniel_hozac, so I come to the conclusion that he is quite busy .... 1147274560 M * derjohn cehteh, ping 1147275241 Q * shedi Quit: Leaving 1147275449 J * comfrey ~comfrey@h-66-167-52-242.sttnwaho.covad.net 1147275460 M * Bertl welcome comfrey! 1147275501 M * comfrey yo B! 1147275679 A * Bertl .o( hmm, so I guess I can start a career as rapper if coding doesn't cut the mustard anymore :) 1147275809 J * azazel ~azazel@81-174-45-117.f5.ngi.it 1147275911 M * Bertl welcome azazel! 1147275937 M * azazel hi Bertl! how are you all? 1147276139 J * id23 ~id@p54A01DAF.dip0.t-ipconnect.de 1147276146 M * id23 back 1147276159 M * Bertl hey id23! 1147276169 M * Bertl azazel: fine, thanks! 1147276722 J * cryo ~say@psoft.user.matrix.farlep.net 1147276923 M * Bertl wb cryo (say)! 1147277064 N * otaku42 otaku42_away 1147277137 Q * pusling Ping timeout: 480 seconds 1147277219 J * f_ ~f_@83-215-237-1.seek.stat.salzburg-online.at 1147277272 M * doener_ daniel_hozac: regarding unwanted .orig files, I've once written a small bash script to help creating kernel diffs... http://www.13thfloor.at/~doener/vserver/tools/kdiff 1147277321 M * doener_ takes 2 or 3 parameters, old-kernel, new-kernel and output file. if no file is given it strips the part before the first dash and replaces it with "patch" 1147277359 M * doener_ ehrm, stripping/replacing from the name of the new kernel 1147277366 J * bonbons ~bonbons@83.222.38.81 1147277387 M * doener_ takes a bit longer than a plain diff because it creates a hard-linked temporary copy, but works well for me 1147277407 M * doener_ it's quite limited, but feel free to improve :) 1147277574 M * Bertl hmm, -x should do nicely, no? 1147277626 M * doener_ I'm just too lazy to do that when make distclean already has all the stuff 1147277629 M * Bertl anyway, off getting some groceries ... back later 1147277650 N * Bertl bertl_oO 1147277657 N * bertl_oO Bertl_oO 1147279332 M * Loki|muh daniel_hozac: hmmm volume removing doesn't work, even when there is only one vserver on that host :( 1147279617 Q * comfrey Ping timeout: 480 seconds 1147279671 M * daniel_hozac doener_: thanks. 1147279692 M * daniel_hozac Loki|muh: hmm, really? so no other guest running? 1147279794 M * Loki|muh no 1147279881 M * daniel_hozac Loki|muh: interesting... 1147280075 M * waldi daniel_hozac: why? 1147280206 M * daniel_hozac waldi: i was just not thinking... somewhy it seemed wrong to me at the time. 1147280260 M * derjohn daniel_hozac, you must be blessed ;) 1147280301 M * daniel_hozac hehe. 1147280344 M * Loki|muh hehe 1147280627 J * Dr4g ~Dr4g@82-40-45-226.cable.ubr06.uddi.blueyonder.co.uk 1147280668 M * daniel_hozac Loki|muh: how do you remove an LV? 1147280800 M * daniel_hozac Loki|muh: got it. i can't reproduce your issue here. 1147280801 Q * mire Ping timeout: 480 seconds 1147280882 Q * Dr4g Quit: 1147280914 J * Dr4g ~Dr4g@tor-irc.dnsbl.oftc.net 1147281107 M * daniel_hozac Loki|muh: so what kernel do you use? utils? 1147281169 M * Loki|muh daniel_hozac: i will try to figure out if there is not a local problem here. I will sneak tomorrow into that. no time today :( 1147281193 M * Loki|muh good to know that it should be something local here ;) 1147281223 M * Loki|muh util-vserver_0.30.209 && 2.6.15.1-vs2.1.0.5 1147281223 M * daniel_hozac well, it could just be that i know nothing of LVM and my test setup was incorrect :) 1147281263 M * daniel_hozac 0.30.210, and 2.6.16-1.2111_FC5.vs2.0.2.0.rc19.1 here. 1147281329 M * Loki|muh oh i missed a release of the tools 1147281449 J * hillct ~H@client200-5.dsl.intrex.net 1147281454 Q * lilalinux Ping timeout: 480 seconds 1147281585 M * Loki|muh someone knows if theres a util-vserver .deb for ubuntu and amd64? 1147281599 M * Loki|muh or debian 1147281777 M * daniel_hozac ftp://ftp.debian.org/debian/pool/main/u/util-vserver/ ? 1147281863 J * mire ~mire@126-167-222-85.COOL.ADSL.VLine.verat.net 1147281873 M * Loki|muh I can't believe that I see there the newest utils... fascinating :) 1147281895 M * daniel_hozac well, micah has been keeping them up to date. 1147281948 M * micah trying my best, sometimes I am overwhelmed and a little slow :) 1147281957 J * lilalinux ~plasma@dslb-084-059-004-250.pools.arcor-ip.net 1147281977 M * Loki|muh micah: great :) 1147282015 M * Loki|muh micah: I assume the latest packages are built for unstable. is this correct? 1147282068 M * daniel_hozac yep. 1147282084 M * daniel_hozac backports.org has sarge builds. 1147282289 J * shedi ~siggi@inferno.lhi.is 1147282314 Q * Dr4g Ping timeout: 480 seconds 1147282614 Q * yarihm Write error: connection closed 1147283264 N * Bertl_oO Bertl 1147283269 M * Bertl back now ... 1147283328 M * hillct Bertl: ! 1147283356 M * Bertl hey hillct! 1147283399 M * hillct Is there any vserver config setting you can think of that would result in network packet loss when changing contexts? 1147283410 M * hillct that also seems to impact the host server 1147283417 M * hillct context 0 1147283515 M * hillct daniel_hozac: suggested that I change the nice value of context 0 processes 1147283634 M * hillct Bertl: the testing done yesterday allowed me to exclude network hardware problems as the cause 1147283767 M * Bertl strange, please elaborate 1147283850 M * hillct well 1147283868 M * hillct I've been seeing periodic massive packet loss 1147283881 M * hillct only occurs when vserver guests are running 1147283908 M * hillct packet loss occurs host only IPs and vserver guest IP 1147283911 M * hillct ips 1147283937 M * hillct I configured the second network interface with your help yestarday to eliminate hardware problems as the issue 1147283981 M * hillct I upgraded from vserver 2.01-pre2 to 2.02 to eliminate issues prior to the release 1147283987 M * Bertl hmm, let's check with ping first 1147283991 J * prae ~benjamin@sherpadown.net 1147284001 M * hillct remaining issue appears to be config related 1147284001 M * Bertl try to ping -f on a local network (but remote unit) 1147284002 M * hillct so 1147284041 M * hillct I don't have a remote unit with which to test, but this was tested earlier, and the behavior was, you'd see 0% packet loss, then you'd see a freeze 1147284054 M * hillct several seconds in durration 1147284063 M * hillct then normal traffic again 1147284076 M * hillct this occurs on both interfaces 1147284101 M * hillct although the second interface still needs to have the vlan configured, I see the behavior when pinging the gateway 1147284163 M * hillct the overall result over a test period of 1 hour or so is between 10 and 30% packet loss 1147284167 M * hillct but it's periodic 1147284177 M * hillct so there'd be a period of 0% loss 1147284184 M * hillct then a few seconds of 100% loss 1147284189 M * hillct then 0% again 1147284198 M * hillct it seems to be vserver load dependant 1147284206 M * hillct as to the frequency and durration of the loss 1147284255 M * hillct while the overall percentages of loss change, the behavior remains the same across all IPs (all vservers and host) 1147284310 M * Bertl welcome prae! good to see you again! 1147284316 M * prae :) 1147284323 M * hillct the visible behavior with regard to websites hosted on each vserver is that more heavily loaded guests get more network time than less loaded guests, and the context change may contribute to the behavior 1147284327 M * prae thx Bertl 1147284341 M * hillct but that's just anacdotal 1147284350 M * Bertl hillct: sounds like some kind of router arp table issue? 1147284362 M * Bertl hillct: maybe you have a loop in your switches/routers? 1147284369 M * hillct this is what I tried to tel lour network guys when I first saw it 1147284380 M * hillct they don't believe it 1147284412 M * hillct since they clain that no one else hosted at the colo has the same issue 1147284427 M * hillct and we've observed this issue on two separate routers 1147284450 M * hillct each of the two NICs on the server is going through a separate router 1147284474 M * hillct that said, the same guy configured both routers 1147284484 M * hillct so it could simply be a config problem 1147284516 M * hillct still, why would my config in two routers be the only one affected 1147284532 M * hillct and it only occurs when vserver guests are running 1147284556 M * hillct which led me to think about locks in context switching 1147284574 M * hillct although I don't know enough about it to form a meaningful theory in this area 1147284617 M * hillct I see shells freeze periodically as well, for a few seconds and recover 1147284624 J * pusling pusling@195.215.29.124 1147284640 M * hillct I first associated the shell freeze behavior with a scheduler problem 1147284647 M * Bertl hillct: I have no idea, but traffic sent from the host, will not be affected by any of the vserver changes 1147284655 Q * pusling Quit: 1147284658 M * hillct but given the emperical network problem, I tried to explain both as network related 1147284667 P * dhansen Leaving 1147284668 J * pusling pusling@195.215.29.124 1147284672 M * Bertl so, what is left there is: hardware issues, mainline networking issues, and router/setup issues 1147284686 M * hillct yeah 1147284701 M * Bertl as other folks are not really reporting those issues 1147284701 M * hillct well, tested two network cards 1147284705 M * hillct saw the same behavior 1147284724 M * Bertl I would rule out the mainline theory, unless you have a very specific kernel 1147284741 M * Bertl this leaves hardware and routing issues 1147284759 M * hillct well, I tested both of the onboard NICs 1147284764 M * hillct both behave the same way 1147284790 M * Bertl what nics are those? what driver is used? 1147284792 M * hillct that doesn't technically rule out server hardware issues but comes close 1147284814 M * daniel_hozac have you tried disabling swap temporarily? 1147284852 M * hillct not disabling 1147284856 M * hillct but swap is never used 1147284870 M * hillct I have 8GB ram in this box 1147284881 M * hillct usage rarely goes above 7 1147284887 M * Bertl hmm, x86_64? 1147284890 M * hillct yes 1147284897 M * hillct NICs are Broadcom NetXtreme BCM5704 Gigabit Ethernet 1147284898 M * Bertl what network drivers? 1147284922 M * Bertl there are two versions, one mainline and one available from the vendor 1147284938 M * Bertl I had some troubles with the mainline ones recently 1147284982 M * Bertl NetXtreme BCM5703X Gigabit here 1147285086 M * hillct I'm not using the vendor specific ones 1147285103 M * hillct the issue does seem to be server load based though 1147285117 M * hillct in that starting a guest adds load 1147285130 M * hillct oh and 1147285132 M * Bertl I'd suggest you give the bcm-8.3.14 a try 1147285133 M * mnemoc hi, there is anything hidden in the web site or somewhere else about using tun/tap for vservers? 1147285166 J * doener ~doener@i5387E225.versanet.de 1147285168 M * hillct this first started appearing when I upgraded from 2.6.11 and vs1.95 to 2.6.13 and vs 2.01-pre2 1147285171 M * Bertl mnemoc: I remember something, but the short version is: you have to create the tun/tap on the host or give network priviledges 1147285210 M * Bertl hillct: what about trying a 2.6.16.15 kernel with bcm drivers and 2.0.2-rc19 (or 20)? 1147285221 M * Bertl wb doener! 1147285227 M * hillct is RC20 just come out? 1147285228 M * mnemoc Bertl: any trick to do that on 'vserver foo start/stop' ? 1147285239 M * doener hey Bertl 1147285243 M * hillct I can redo the kernel to test, sure 1147285246 M * doener working on the lease stuff now... 1147285255 M * hillct can't do such tests in the middle of the day though 1147285280 M * Bertl doener: please coordinate with daniel_hozac :) 1147285295 M * doener daniel_hozac: so you're also working on that? 1147285330 M * Bertl I will clean up the 2.0.2 and 2.1.1 releases today and tomorrow (and check the code once again) 1147285353 M * Bertl would be great to have proper patches tomorrow to get them into -rc20 1147285369 M * Bertl will also look into reducing the vs_cvirt includes 1147285411 M * doener daniel_hozac: I'm currently trying to use the fl passed to __setlease as the one being inserted. Problem is that we somehow need to know whether it was inserted or not. In the former case, everythings fine. The latter case forces us to free the allocated lease and I'm searching for some nice semantics for __setlease to tell the callee about it 1147285442 M * doener oops, s/callee/caller/ 1147285526 M * hillct Bertl: I'll do another kernel rebuild tonight and see if that improves things 1147285546 M * doener daniel_hozac: I currently have a &fl != flp check, but that's rather ugly, especially as the "new" flp won't be used anyway 1147285573 M * hillct Bertl: I'm currently using the tg3 drivers 1147285658 M * doener hillct: what issues do you see? 1147285711 M * hillct periods of 100% packet loss on host and guest IPs followed by periods of 0% packet loss 1147285747 M * hillct over an extended period the periods of esentially network disconnect account for between 10 and 30% 1147285803 M * hillct according to the Broadcom website, the tg3 drivers are now the officially supported drivers 1147285817 M * hillct I wonder when they were added to the kernel mainline 1147285829 M * hillct Guessing between 2.6.11 and 2.6.13 releases 1147285857 A * hillct will go back to the bcm drivers for testing as bertl suggests 1147285902 M * hillct the more I think about it, the more probable a solution that seems given the timeline of the issue 1147286920 M * daniel_hozac doener: hmm? 1147286941 M * doener could you be a little more specific? ;) 1147286975 Q * FireEgl Ping timeout: 480 seconds 1147286986 M * daniel_hozac do the leases have more problems? or are you just optimizing? 1147287022 J * sid3wind1 luser@bastard-operator.from-hell.be 1147287027 M * Bertl welcome sid3wind1! 1147287028 J * sladen_ paul@starsky.19inch.net 1147287029 M * doener the latter... Bertl didn't fully agree on the file lock on the stack and I don't really like the whole thing either 1147287034 M * Bertl wb sladen_! 1147287132 Q * sid3windr Ping timeout: 480 seconds 1147287144 Q * sladen Ping timeout: 480 seconds 1147287157 Q * kilian Ping timeout: 480 seconds 1147287427 M * sladen_ grettings Bertl 1147287449 J * kilian kk@projects.verfaction.de 1147287571 M * daniel_hozac Bertl: reducing vs_cvirt includes? by the utsname changes? 1147287677 M * daniel_hozac doener: yeah, it's pretty ugly... 1147287725 M * Bertl daniel_hozac: yes, something in that direction 1147287807 Q * lilalinux Remote host closed the connection 1147287830 J * vreak ~irc@83-215-140-176.seek.dyn.salzburg-online.at 1147287842 M * daniel_hozac doener: although, as it's mostly for matching having it on the stack seems like the most sensible thing to do. 1147287902 M * Bertl welcome vreak! 1147288471 Q * mire Ping timeout: 480 seconds 1147288566 M * doener daniel_hozac: I'm trying to avoid that thing altogether 1147288651 M * doener __setlease now allocates a new lock from slab, I'm trying to make it simply insert the lock being passed to it. moving the slab allocations out of __setlease, out of the lock and avoiding the second lock on the stack 1147288688 M * doener the allocated lock has all its properties copied from the lock passed into __setlease anyway 1147288712 M * daniel_hozac yeah. 1147288841 M * doener The simplest thing to do would be returning -EEXIST from __setlease when my_before != NULL, but that's ugly as that error code must not be propagated then 1147288934 M * daniel_hozac and returning -EEXIST on success would be odd. 1147288944 M * doener that as well 1147288989 M * doener I thinking about EPTR stuff now... maybe that works nicely 1147289100 N * sladen_ sladen 1147289155 J * mire ~mire@118-166-222-85.COOL.ADSL.VLine.Verat.NET 1147290298 M * Bertl okay, off for now .. back later ... 1147290304 N * Bertl Bertl_oO 1147290780 J * FireEgl Atlantica@Atlantica.DollarDNS.Net 1147291123 J * yarihm ~yarihm@217-162-112-205.dclient.hispeed.ch 1147291488 M * hillct wow 1147291514 M * hillct This is NOT the sort of build that I'd call confidence inspiring when it comes to network drivers: 1147291516 M * hillct http://pastebin.com/710127 1147292285 J * harti ~Hartmut@85-124-100-166.dynamic.xdsl-line.inode.at 1147292428 Q * shedi Quit: Leaving 1147293147 Q * harti Quit: Client exiting 1147293286 M * derjohn did anyone run a gameserver (CS) within a guest? I have complaints that the thingy lags (but cpu usage is < 60%) , it may have to do with newer kernels. Is there a way to make a single process gain more prio (expect nice ... it did that already) ? 1147293656 Q * f_ Quit: This computer has gone to sleep 1147293701 J * f_ ~f_@83-215-237-1.seek.stat.salzburg-online.at 1147293833 Q * f_ Quit: 1147294308 J * f_ ~f_@83-215-237-1.seek.stat.salzburg-online.at 1147294527 J * doener__ ~doener@i5387D6B1.versanet.de 1147294545 M * derjohn is there a particular prob with "chrt" and a vserver enabled kernel? 1147294559 M * derjohn hy, doener .. now I get hungry ;) 1147294566 M * derjohn doener, ip a :) ! 1147294675 M * doener__ :) 1147294699 M * doener__ my underscore level increases ;) 1147294705 M * derjohn __________cool ! 1147294751 Q * doener Ping timeout: 480 seconds 1147294768 Q * doener_ Ping timeout: 480 seconds 1147294800 M * derjohn do you know if the scheduler behaves bitchy compared to a non vserver kernel? (for sensitive apps like voip oder gameservers) ? 1147294985 M * doener__ no idea 1147295306 M * derjohn CS sucks in a guest, but also on the same host ... and Bertl_oO fingered around in the scheduler since rc16+ or so... 1147295335 N * Bertl_oO Bertl 1147295344 M * Bertl did somebody talk about me? 1147295360 M * derjohn Hey, the Bertl is back ;) 1147295366 M * derjohn morning ! 1147295368 M * Bertl derjohn: the scheduler is the same as in mainline 1147295377 M * doener__ "Wenn man vom Teufel spricht..." ;) (dunno if there's an english version of this...) 1147295392 M * Bertl as long as you do not activate the hard cpu scheduler, nothing will change 1147295399 M * derjohn i just used "chrt" (without knowing anything about rt linux). the box froze. 1147295433 M * derjohn activate means: set the kernel option at compile time? of activate within at least one guest? 1147295444 M * Bertl a) what is chrt supposed to do, and b) where did you run it? 1147295468 M * Bertl sounds to me like setting a realtime priority for a task 1147295501 M * derjohn a) set the realtime (thus rt) scheuler and prio. there is RR, FIFO and OTHER (which is default) and a numeric value from 1 to 99. 1147295537 M * derjohn b.) on the host. 2.6.16.14 vs 2.1.1 rc 19 @ 250 HZ. 1147295539 M * doener__ and you set RR or FIFO? that will keep the process running until it decides to end or block 1147295567 M * doener__ (well, or it can yield the cpu, but that'll only have an effect if there's more than one rt process IIRC) 1147295581 M * derjohn doener__, I tried fifo .... for the particular gameserver pids ... (so re-sched, not a fresh start) 1147295598 M * derjohn box box froze, the remote hands are on their way :) 1147295648 M * doener__ hm, I'd not expect it to freeze, just to schedule nothing but that gameserver 1147295684 M * Bertl derjohn: did the game server stop working? 1147295738 M * derjohn doener__, agreed. Bertl, well the whole machine froze. now I press thumbs that the fsck is fine. 1147295753 M * Bertl hmm? no magic sysrq? 1147295796 M * Bertl derjohn: well, it should not freez, only shedule only that process 1147295839 J * harti ~Hartmut@85-124-100-166.dynamic.xdsl-line.inode.at 1147295891 Q * f_ Quit: Leaving 1147295896 M * derjohn magic sysrq? I have remote hands ;) ... (and actually ext3 didnt make much problems in the last 2 years ...) 1147295917 M * Skram hi all 1147295931 M * derjohn I think the remote fingers hit the wrong button ("push the button baby ...") ... 1147296212 J * comfrey ~comfrey@h-64-105-215-74.sttnwaho.covad.net 1147296233 Q * prae Quit: Pwet 1147296431 M * Radiance evening all :) 1147296456 M * Radiance when i backup a vserver config and data, upgrade to a new kernel and new vserver patch, i can deploy the backup without much change ? 1147296623 M * Bertl evening Radiance! 1147296656 M * Radiance hiya mate, how are things today ? :) 1147296658 M * Skram should. 1147296661 M * derjohn Bertl, the machine is up again. Silly typo in /etc/network/interfaces. :( 1147296681 M * Bertl Radiance: fine, tx! 1147296725 M * derjohn is there a way to give a certain process more power? I mean, more like with nice. thats what brought mw to the chrt idea .. 1147296771 M * derjohn the gameserver lags, but only consumes < 60% cpu. There is nearly no I/O and the bandwidth is ok (other machines dont lag) 1147296793 M * Bertl well, nice _is_ the tool for that 1147296935 M * hillct oh what fun 1147296947 M * Skram is 77 degrees farenheit a good temperature for a 1U dual xeon server? What is the max it should ever reach (not safe)? 1147296963 M * hillct not only is the bcm5700 driver not supported, but it won't compile for any kernel after .13 1147297030 M * cehteh Skram: Netburst based Xeons? they limit/throttle themself automatically safely ... but having proper cooling is always good 1147297034 M * hillct TONS of patches to the tg3 driver code since Feduary, all from the Broadcom guys 1147297037 M * hillct :( 1147297047 Q * softi42 Ping timeout: 480 seconds 1147297062 M * Skram cehteh: two fans, i think is in them 1147297070 M * Skram cehteh: I dont knwo what netburst is 1147297072 M * Skram *shrug* 1147297083 M * cehteh i dont know how much celsius that is 1147297104 M * cehteh current ones based on the same core as the P4 1147297168 M * cehteh well al long you are over the dew-point you cant cool too much :) .. only to expensive 1147297201 M * Skram ok 1147297203 M * Skram thnx. 1147297272 M * cehteh (better read the datasheet for that machine to get exact specs) ... hardisks and capacitors are less heat-toleant than just the CPU's 1147297286 M * Skram yeah 1147297290 M * Skram thanks again. 1147297310 M * cehteh there is a hddtemp daemon/command for linux 1147297415 M * derjohn Bertl, is niced the process to -5 and it got slower (!) (more laggy). 1147297465 M * Bertl derjohn: what is the priority of that process? 1147297492 M * derjohn cehteh, that doesnt work for sata (last time i checked) 1147297508 M * cehteh derjohn: the kernel has some race-conditions ... i had the same with encrypted devices .. niceing (positive) the kcryptd actually improved performance 1147297520 M * derjohn Bertl, it was 0 I niced it to -5 ... or what kind of prio are you talkung about0? 1147297539 M * cehteh depends on the chipset .. i have here 4 identical hdds on 2 controlers 2 work, 2 not 1147297542 M * derjohn cehteh, somtimes i hate computers 1147297564 M * cehteh derjohn: well profile that thing and check where the performance get sucked 1147297573 M * cehteh database? 1147297573 M * derjohn cehteh, hddtemp? or are you talking about lm/acpi temp ? 1147297574 M * Bertl PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 1147297575 M * Bertl 1 root 16 0 1564 540 472 S 0.0 0.0 0:01.31 init 1147297584 J * softi42 ~softi@p549D77E0.dip.t-dialin.net 1147297586 M * Bertl derjohn: see the PR and NI columns? 1147297596 Q * dna Quit: Verlassend 1147297610 M * cehteh starbase:~# hddtemp /dev/hdg 1147297610 M * cehteh /dev/hdg: Maxtor 6B200M0: 32 C 1147297615 M * cehteh .. thats what i mean 1147297646 M * derjohn Bertl: 6250 silvercl 15 0 86172 66m 9224 S 3.3 7.0 0:17.22 hlds_amd -> 15 and 0 1147297680 M * derjohn But this is now on a non-VS 2.6.16, K7 optimized. ( the other was a p3 with generic x86 opt ...) 1147297687 M * Bertl so it is not reniced atm 1147297695 M * derjohn the machine is an athlon 64 3000+ 1147297698 M * cehteh derjohn: makeing something realtime can mean that it had to do more/smaller IO request since its scheduled faster/more often 1147297713 M * cehteh i wont recommend realtime for any process which does io 1147297744 M * derjohn Bertl, no, it's rebooted and started fresh. I wait for user feedback if the non-vs kernel is more responice. 1147297777 M * derjohn Bertl, I heared that something in nice was tweaked in vs? I.e. setting smaller values is not possible in a guest? 1147297796 M * cehteh that needs CAP_SYS_ADMIN 1147297801 M * Bertl derjohn: you can always do that from the host 1147297801 M * cehteh (i think) 1147297808 M * derjohn cehteh, the is CAP_NICE AFAIR 1147297812 M * cehteh maybe 1147297874 M * cehteh anyways ... just setting something less nice will be actually just the wrong thing if the cause of some latency is somewhere else (io elevator, processes which have to serve some data etc) 1147297956 Q * bonbons Quit: Leaving 1147298637 Q * yarihm Quit: Leaving 1147299305 M * Skram cehteh: . 1147299306 M * Skram hercules / # hddtemp /dev/sda 1147299307 M * Skram /dev/sda: ATA WDC WD1600JS-75M: drive supported, but it doesn't have a temperature sensor. 1147299310 M * Skram oh well. 1147299798 M * derjohn Skram, WD sucks. 1147299889 Q * vreak Quit: ( www.nnscript.de :: NoNameScript 3.81 :: www.XLhost.de ) 1147299928 M * Skram yeah 1147300881 M * cehteh haha 1147300930 M * cehteh all harddrives are bad ... WD is just a bit more bad :) 1147301092 Q * comfrey Ping timeout: 480 seconds 1147301654 M * Skram haha 1147302930 Q * azazel Quit: Client exiting 1147303146 J * after ~after@200.96.91.159 1147303163 M * after Can I create a vserver under redhat? 1147303188 M * derjohn after, yes. 1147303199 M * after where can I find a tutorial about it? 1147303248 M * daniel_hozac vserver - build --help 1147303277 M * after i need to recompile my kernel or something like this ? 1147303326 M * daniel_hozac yep, of course. 1147303358 M * derjohn daniel_hozac, arent you a redhat fan? 1147303360 M * daniel_hozac without further specifying your distribution, http://linux-vserver.org/Step-by-Step+Guide+2.6 is generally applicable to all. 1147303368 M * after ok 1147303406 M * derjohn daniel_hozac, dont fc tools work for redhat? 1147303409 M * daniel_hozac you might want to consider 2.6.16.15 and 2.0.2-rc19 rather than the versions used in that document. 1147303416 M * daniel_hozac it depends on what "redhat" means. 1147303442 M * daniel_hozac hence the "without further specifying your distribution" above ;) 1147303455 M * derjohn after, if there is a change take daniels tools. they .. just work. 1147303467 M * after hehe ok 1147303516 Q * mire Ping timeout: 480 seconds 1147303619 Q * after Quit: Conecte-se por /server irc.viairc.com.br 1147303777 Q * FireEgl Quit: Bye... 1147304186 J * mire ~mire@200-167-222-85.COOL.ADSL.VLine.Verat.NET 1147304414 J * Oli ~skycode@212.224.224.150 1147304781 Q * Oli Quit: Oli