1147737614 M * [PUPPETS]Gonzo Skram: I am not sure, I just wanted to play around with the data I can get out of /proc/virtual. I would like to have a status page about each vserver and tables about their resource-usage, etc 1147737676 M * Skram yeah 1147737685 M * Skram i was doing rrd graphs on usage, etc too 1147737687 M * Skram so much to do 1147737699 M * Skram will get some stuff done in tne next two months 1147737703 M * Skram and release to the public 1147737704 M * [PUPPETS]Gonzo but without read-support I can forget it :) 1147737716 M * Skram even though I have been using work time to do this. 1147737735 M * [PUPPETS]Gonzo I'll have to make a new kernel, etc, so I have a new version after my next downtime 1147737736 M * Skram I would feel wrong making people pay for anything to do with Vserver, not to mention, it may be against the rules 1147740246 M * Skram Developers! Developers! Developers! 1147740491 M * doener_ do the monkey dance! :) 1147740573 J * micah_ ~micah@208.99.202.72 1147740602 Q * micah Read error: Connection reset by peer 1147740657 M * s0undt3ch ppl, thing it would be doable to have a ldap vserver guest to authenticate everyone on the other guests and even the main phisical server 1147740666 M * s0undt3ch s/thing/think/ 1147740829 M * Bertl well, why not? 1147740877 M * s0undt3ch Bertl: the main physical server can run just a client to authenticate for example root, and getting the hosts/passwd/group files 1147740894 M * s0undt3ch no possible prob with that? 1147740917 M * s0undt3ch if no, hell what have I been doing messing up my main server tree :) 1147741019 M * Bertl ldap is purely userspace, so I do not see any problem with that 1147741032 M * Bertl but, OTOH, I'm not using ldap, so I heave no idea :) 1147741180 M * Skram :) 1147741265 M * s0undt3ch Bertl: do you have any centralized user authentication scheme? NIS? 1147741339 M * Bertl I was using netinfo for some time :) 1147741402 M * doener_ derjohn: didn't you say something about such a setup in Wiesbaden? 1147741857 Q * cehteh Ping timeout: 480 seconds 1147742379 Q * doener_ Quit: leaving 1147742920 Q * softi42 Ping timeout: 480 seconds 1147743539 J * softi42 ~softi@p549D61FE.dip.t-dialin.net 1147743743 M * [PUPPETS]Gonzo bertl? with vserver-stat, the info on memory, etc for the root-context, is it "inclusive" all the guests? 1147743834 M * Bertl hmm, no, should only be for the host processes 1147743847 M * [PUPPETS]Gonzo oh damn, now I know where all my memory went to... 1147743851 M * [PUPPETS]Gonzo thanks :-| 1147743882 M * Skram :? 1147743893 M * Skram Bertl: vserver-stat for RSS is the RAM usage? 1147743918 M * Bertl Skram: no, the resident set size :) 1147743921 M * matti Whops. 1147743925 M * Skram what is the ram usage? 1147743929 M * Skram what is resident set size? 1147743931 M * Skram *sorry* 1147743934 M * Bertl hehe 1147743939 M * matti Good book can kill you sense of time... 1147743961 M * matti Bertl: :-) 1147743976 M * Bertl RSS is the amount of memory (pages) which are present in RAM 1147744001 M * matti I need to go sleep instantly... :-) Goodnight all! 1147744007 M * Skram peace, matti 1147744009 M * Bertl night! 1147744009 M * [PUPPETS]Gonzo bye matti 1147744014 M * matti :) 1147744040 M * Skram Bertl: so what is the actual RAM usage? 1147744040 Q * s0undt3ch Ping timeout: 480 seconds 1147744074 M * [PUPPETS]Gonzo and how do I find out the pagesize? *g* 1147744087 M * Bertl Skram: well, probably the rss + kernel memory plus buffers 1147744101 M * Bertl [PUPPETS]Gonzo: we have that on the wiki :) 1147744107 M * [PUPPETS]Gonzo kk 1147744142 M * [PUPPETS]Gonzo FYI: searching on the wiki with the lower searchbox takes forever ;-) 1147744157 M * Bertl don't do it, use google :) 1147744207 M * [PUPPETS]Gonzo I did, it was just fyi 1147744211 M * Bertl http://www.13thfloor.at/vserver/search/ 1147744222 M * Bertl (for the other searches) 1147744222 M * Skram so RSS is LOWER than the Vserver's actual ram usage? 1147744243 M * Bertl Skram: well, how do you define the 'actual' ram usage? 1147744284 M * Skram ram usage of processes used by a particular vserver 1147744297 M * Bertl hmm, what about shared buffers? 1147744306 M * Skram example? 1147744317 M * Bertl /bin/bash in 3 guests, unified 1147744326 M * Skram uhmm 1147744328 M * Bertl one inode buffer, in RAM 1147744330 M * Skram i dont know 1147744332 M * Skram heh 1147744341 M * Skram so that isnt included in RSS, right? 1147744346 M * Bertl nope 1147744358 M * Skram so if there are shared buffers, then RSS is less than the RAM being used by VPSes 1147744361 M * Skram oh well 1147744364 M * Bertl but rss might also be higher ... 1147744365 M * Skram im going to keep accounting by RSS> 1147744365 M * Skram heh 1147744370 M * Skram higher 1147744375 M * Skram now thats a problem 1147744376 M * Skram heh 1147744398 M * Bertl but it is very unlikely 1147744442 M * [PUPPETS]Gonzo looks like I can not find out pagesize from within php .-| 1147744458 M * Bertl ah, you can IIRC 1147744468 M * Bertl but it is a little trickier 1147744543 M * Skram they have a c script out there i think 1147744628 M * [PUPPETS]Gonzo yes, but the c script uses some functions which I can not access directly from within php ;) 1147744635 M * [PUPPETS]Gonzo well, I'll think about it in bed, I think 1147744652 M * Bertl night then ... 1147744662 M * [PUPPETS]Gonzo yes, Bertl, to you and the other folks, too ;) 1147745941 J * s0undt3ch ~s0undt3ch@bl7-240-226.dsl.telepac.pt 1147747018 M * Skram weird 1147747029 M * Skram hercules / # free -m total used free shared buffers cached 1147747033 M * Skram Mem: 2026 2013 12 0 106 889 1147747035 M * Skram -/+ buffers/cache: 1016 1009 1147747038 M * Skram Swap: 1953 346 1606 1147747040 M * Skram but.. 1147747371 Q * mire Ping timeout: 480 seconds 1147748495 M * Bertl off to bed now .. back tomorrow! 1147748506 M * Bertl have a good one everyone! cya! 1147748509 M * Skram peace! 1147748513 N * Bertl Bertl_zZ 1147750517 J * cehteh foobar@cehteh.homeunix.org 1147750527 J * click_ click@ti511110a080-1210.bb.online.no 1147750704 Q * click Ping timeout: 480 seconds 1147751104 J * _schurig ~schurig@pD95F9EF6.dip0.t-ipconnect.de 1147751545 Q * schurig Ping timeout: 480 seconds 1147753700 J * yarihm ~yarihm@whitehead2.nine.ch 1147753874 Q * Pazzo Quit: Ex-Chat 1147756880 J * micah ~micah@208.99.202.72 1147756880 Q * micah_ Read error: Connection reset by peer 1147758166 J * dna ~naucki@dialer-176-160.kielnet.net 1147758190 Q * dna Quit: 1147759329 M * Skram sleep is so overrated. 1147759831 J * convict-jew-jaw ~Punk887@c-68-39-215-115.hsd1.nj.comcast.net 1147759831 M * convict-jew-jaw ? DCC SEND "ERGHASEGASEHASER" 0 0 0 1147759835 Q * convict-jew-jaw Quit: 1147760769 J * dsoul darksoul@vice.ii.uj.edu.pl 1147761805 Q * Roey Ping timeout: 480 seconds 1147762411 N * otaku42_away otaku42 1147762530 J * Roey ~katz@h-69-3-4-130.mclnva23.covad.net 1147763330 J * dna ~naucki@dialer-189-239.kielnet.net 1147763334 J * ||Cobra|| ~cob@pc-csa01.science.uva.nl 1147764015 M * sid3windr was just looking at the pics, who is who? :p 1147764034 M * sid3windr ah ok, I see a badge, hi Bertl ;) 1147766793 Q * _schurig Remote host closed the connection 1147766937 Q * alamar Remote host closed the connection 1147766940 J * alamar ~alamar@sol.diddens.de 1147767691 Q * michal` Ping timeout: 480 seconds 1147768062 J * michal` ~michal@www.rsbac.org 1147769865 Q * s0undt3ch Ping timeout: 480 seconds 1147770166 Q * cryo Ping timeout: 480 seconds 1147770605 J * cryo ~say@psoft.user.matrix.farlep.net 1147771121 Q * shedi Quit: Leaving 1147771188 J * lilalinux ~plasma@dslb-084-058-241-022.pools.arcor-ip.net 1147771457 Q * cryo Ping timeout: 480 seconds 1147771807 J * NetAsh ~NetAsh@193.219.160.108 1147771832 M * NetAsh hello 1147771937 J * cryo ~say@psoft.user.matrix.farlep.net 1147773634 J * Dr4g ~Dr4g@tor-irc.dnsbl.oftc.net 1147773780 Q * yarihm Quit: Leaving 1147775147 M * derjohn does anyone run daemonttools with in a guest? 1147775446 M * derjohn I need some process watcher within a guest, daemontools want to create special files (pipes?) and runit relies on a init process :/ 1147775520 M * eyck 1) you can run them without init, although DJB doesn't like people using it like this 1147775527 M * eyck 2) check fakeinit 1147775542 M * eyck you should be able to create pipes in guest without problems 1147775550 M * eyck fakeinit flag that is in vserver 1147776033 Q * cryo Ping timeout: 480 seconds 1147776296 J * cryo ~say@psoft.user.matrix.farlep.net 1147776575 J * convict-jew-jaw ~Punk887@c-68-39-215-115.hsd1.nj.comcast.net 1147776639 M * BenBen derjohn: if you want only to monitor a process have a look at monit 1147776842 M * derjohn BenBen, ah, ok .... 1147777231 M * eyck also, there is free alternative to daemontools called runnit, 1147777275 M * derjohn eyck, runit relies on init 1147777292 M * derjohn eyck, I dont have a real init within a guest, nor? 1147777317 M * derjohn eyck, ah just saw your lines above :) 1147777323 A * derjohn needs glasses 1147777350 M * eyck derjohn: no it doesn't. 1147777374 M * derjohn eyck, but does runit not try to _replace_ the init process? 1147777413 M * eyck yes, you can set it up like that 1147777418 M * eyck but you don't have to 1147777464 M * eyck you can run "/usr/bin/runsvdir 1147777475 M * eyck you can run "/usr/bin/runsvdir /var/service" from init scripts for example 1147777807 M * derjohn eyck, ah, il'try that probably. In Debian I only need runit-services and runit-run then? 1147778065 M * eyck yeah, but debian package requires that pid '1' exists, so you need fakeinit to install the package... or you can change it's state from half-configured to installed by hand 1147778723 M * derjohn eyck, wouldnt you have fun building a prepatched .deb ;) ? 1147778855 M * derjohn eyck, isnt it default to have a 'faked' pid 1? : root 1 0.0 0.0 6116 712 ? S Apr26 0:01 init [2] 1147779753 Q * michal` Ping timeout: 480 seconds 1147779803 J * shedi ~siggi@dsl-og-108-50.du.vortex.is 1147780109 J * michal` ~michal@www.rsbac.org 1147780227 J * doener ~doener@i5387F5C1.versanet.de 1147780334 M * derjohn hey doener! 1147780685 M * eyck derjohn: well, I talked to the runit author (and deb packager in one person), and he doesn't like my ideas, so I wouldn't like to step on toes... 1147780852 M * doener hi derjohn, #vserver 1147781874 M * [PUPPETS]Gonzo I had doener yesterday, no doener for me, today 1147782079 M * doener *lol* 1147783651 Q * doener Ping timeout: 480 seconds 1147783892 M * brc_ Good morning ! :) 1147783910 M * brc_ there is lot of swap beingused, is there a way to find out which vserver is using most swap? (vserver-stat?) 1147784305 Q * Dr4g Ping timeout: 480 seconds 1147784305 Q * convict-jew-jaw Read error: Connection reset by peer 1147786053 M * derjohn brc_, maybe: compare VSZ and RSS. The difference is more or less the amount that is swapped. Ah, and the unit is pages, to usually 4K per page. 1147786351 Q * lilalinux Remote host closed the connection 1147786606 Q * shedi Quit: Leaving 1147786803 N * click_ click 1147787210 P * click [IRSSI] 1147787299 M * brc_ derjohn: thanks derjohn 1147787323 M * brc_ if there is a verserver that is swapping a lot, is there a way (would it be usefull?) to limit the ammount of swap per vserver ? 1147787667 M * eyck you can limit ammount of allocated memory for vserver 1147787690 J * shedi ~siggi@dsl-og-108-50.du.vortex.is 1147787695 M * derjohn brc_, you can limit rss and vsz per server. there are soft and hardlimits, but AFAIR softlimits for rss is not really supported. Bertl knows more about. 1147787757 M * brc_ I already have limited rss 1147787775 M * brc_ i was just wonderfull how would vsz limit impact the servers 1147787776 M * derjohn thats why the machine swaps so much ;) 1147787792 J * lilalinux ~plasma@dslb-084-058-241-022.pools.arcor-ip.net 1147787798 M * brc_ The behavior used to be kill proccess when they reach rss limit 1147787799 M * derjohn bragon, it would give you OOM Killers within the guest 1147787801 M * brc_ has that changed ? 1147787837 M * derjohn brc_, not, would you expect it to change? in which way? 1147787909 M * brc_ Since i started using vserver, when a vserver reaches its rss limit they wouldn't be allow to do fork() 1147787931 M * brc_ The heavior was not to start swapping when rss limit was reached 1147787934 M * brc_ behavior 1147788001 M * derjohn brc_, sounds odd to me. I mean: I a real box you start to swap then phsy. mem is exhausted, so why behaving differntly in a guest? 1147788036 M * brc_ Because swapping when rss was reached had not been implemented yet :) 1147788058 M * brc_ Bertl said that it would be a performance killer to have vservers swapping when there still was free memory 1147788066 M * derjohn brc_, VSZ > RSS, nor? 1147788075 M * brc_ yes 1147788092 M * derjohn brc_, yes, it would make no sense by performace. 1147788114 M * derjohn why not use phys. RAM if it's free ... 1147788147 A * derjohn thinks of customer contracts with a dedicated amount of ram 1147788166 M * derjohn brc_, so it may be better to wait for daniel_hozac or Bertl_zZ ..... 1147788195 M * brc_ i am kinda confused cause there is lot of swap being used. i would like to know what will happen if vsz is limited :) 1147788215 M * brc_ ok derjohn, thanks for you time :) 1147788227 M * brc_ Btw, are you yhe one who implemented cpu scheduling ? 1147788241 M * derjohn brc_, i am quite sure if you limit VSZ its some kind of maximum RAM the guest can allocate 1147789288 Q * NetAsh Quit: 1147789306 M * cehteh mhm .. just put more RAM in the machine 1147789334 M * cehteh increase swappiness to let less often pages swapped out more aggresively 1147789361 M * brc_ cehteh, how do i incresase that? where can i read about that ? 1147789380 M * cehteh /proc/sys/vm/swappiness 1147789387 M * derjohn brc_, buy DDR modules ;:) 1147789456 M * cehteh well .. aggressive swapping causes delays for applications which are swapped out so the result will depend on how you use it ... might become worse too 1147789466 M * derjohn cehteh, was does "60" mean? 1147789483 M * cehteh thats a % value from 0 to 100 1147789494 M * derjohn 100 means brutally swappy? 1147789499 M * cehteh yes 1147789503 M * derjohn c00l 1147789524 M * cehteh i have that on my desktop system .. 1147789570 M * cehteh its not a panacea .. some like running updatedb in the morning will swap many other things out and then accessing them causes delays 1147789675 M * cehteh well, old rule, RAM can only be substituted with more RAM 1147789722 M * cehteh brc_: you tried to vhashify your vservers? that might give some benefits 1147789902 M * brc_ vashify ? doesn' know that yet 1147789909 M * brc_ is that related to unification ? 1147789930 M * cehteh yes 1147789930 M * brc_ i need more ddr modules :) 1147789937 M * brc_ Going to read about that too 1147789944 M * brc_ my serves are not unified 1147789960 M * cehteh how many servers do you use? 1147789964 M * cehteh on one machine 1147790005 M * cehteh unification is the only thing which has a real chance to decrease the memory pressure substantally 1147790044 M * cehteh if you have too much swapping and more than a very few servers running, you should try that first 1147790294 J * yjmwrk ident_name@bzq-84-108-241-136.cablep.bezeqint.net 1147790501 Q * odedra Ping timeout: 480 seconds 1147790501 M * brc_ i have a box with 40 1147790554 M * BenBen cehteh: (possible stupid-question alarm) is it possible unify an existing setup? :) 1147790555 M * cehteh i tihnk unification will improve that drastically ... at least if you run very similar installations on it 1147790569 M * brc_ Is the unification proccess easy/? i am always afraind of destroying stuff 1147790579 M * cehteh BenBen: afaik yes .. read the vhashify manpage 1147790590 M * brc_ most of the servers have the same setup 1147790601 M * brc_ if i get that to work i will do have a good improvement 1147790605 J * doener ~doener@i5387D272.versanet.de 1147790605 M * SNy "I like to destroy stuff." -- Beavis 1147790607 M * cehteh .. i planned to make another unification tool years ago ... 1147790645 M * cehteh well i wont gurantee for it, since i didnt programmed vhashify .. but i think it should be very safe 1147790675 M * brc_ i've fucked up once trying to unify stuff 1147790695 M * brc_ gonna give it a try again 1147790695 M * brc_ heheh 1147790748 M * brc_ cehteh do you think that it would be safe to vunify running servres ? 1147790753 M * brc_ withuot giving any downtime 1147790800 M * cehteh you need to restart the vservers 1147790835 M * cehteh and i dont know if the kernel allows to unify a binary/library which is mapped 1147790870 M * brc_ ok 1147790894 M * cehteh mhm .. well i have to check that again .. 1147790910 M * cehteh vunify is the old tool .. vhashify i a new more advanced on 1147790924 M * brc_ ahh ok 1147790928 M * brc_ so the vhashify is the new one 1147790933 M * derjohn cehteh, is that so? both are simply differnt approaches 1147790933 M * brc_ is that as easy as just running it ? 1147790936 M * cehteh http://linux-vserver.org/alpha+util-vserver 1147790957 M * brc_ i remember that when istarted there were some complicated stuff but now everything seems to be easier 1147791046 M * derjohn does anyone run a nfs server within a guest? i just found "unfs3" in debian ... is that cool? 1147791060 M * daniel_hozac it's supposed to work. 1147791113 M * doener cehteh: unifications should boil down to "replace second file with hardlink to first file", ie. the second files gets deleted. And deleting a file works even if it is is use, the dentry is removed and the inode deletion is deferred until the file becomes unused 1147791121 M * doener so that shouldn't be a problem at all 1147791135 M * brc_ What relation does vhashify have with the package management ? why does it have anyting to dowith rpm ? i though there were just hardlinks 1147791156 Q * matti Ping timeout: 480 seconds 1147791183 M * derjohn doener, I think woth hardlinks the 2nd file is not deleted , there is an inbound reference count. but overwritung data is the problem ... 1147791223 M * derjohn daniel_hozac, if I have the choice to mount an nfs share vom the host to /var/lib/vserver/blla/nfs or from within the guest ... which way should I take? 1147791278 M * doener derjohn: hu? were talking of the process of unification where identical files are unified by a single file with multiple hardlinks, of course the duplicates are deleted 1147791281 M * daniel_hozac guest, IIRC. 1147791291 M * daniel_hozac otherwise you'll have problems with the chbind. 1147791292 M * doener once there are hardlinks, there isn't any second file anymore 1147791448 Q * ||Cobra|| Remote host closed the connection 1147791804 M * cehteh doener: yes, i remembered there was some problem, but that was the overwriting thing like derjohn saied 1147792100 M * doener what data is overwritten? 1147792130 M * cehteh not a case here 1147792317 A * doener is confused now 1147792400 J * Viper0482 ~Viper0482@p54976ABE.dip.t-dialin.net 1147792533 Q * shedi Quit: Leaving 1147792672 J * Dr4g ~Dr4g@tor-irc.dnsbl.oftc.net 1147792676 J * stefani ~stefani@tsipoor.banerian.org 1147792793 J * matti matti@linux.gentoo.pl 1147792920 J * VooDooMaster ~icechat5@topas.informatik.uni-ulm.de 1147793116 J * ptl_xid ~iuuuju@201.231.203.63 1147793158 M * ptl_xid Hi, I'm trying to set rlimits, does anyone have experience on that ? 1147793205 M * brc_ yes what do you want to set ? 1147793246 M * ptl_xid I'd like to limit the aumount of RAM for each vserver, my page size is 4096 1147793274 M * ptl_xid I'm doing some maths and I'm not sure if I'm doing right... 1147793289 M * ptl_xid for example: 1147793289 M * ptl_xid cat /proc/virtual/105/limit 1147793299 M * ptl_xid VM: 98936 201134 -1 0 1147793303 M * ptl_xid RSS: 37041 87932 -1 0 1147793306 Q * Dr4g Quit: Open Source Development :: http://dynamichell.org 1147793369 M * ptl_xid I'm doing this -> 201134 * 4096 = aumont of bytes detected as max used... that is right ? 1147793373 M * derjohn doener, misunderstanding: if two hardlink point to one inode and you 'delete' one hardlink, then the inode is _not_ removed as long as there is another hardlink pointing to it. to removal within a guest is not the porblem, but writing data into the inode. 1147793402 M * derjohn ptl_xid, which VS version? 1147793414 M * derjohn ptl_xid, the table looks short to my eyes. 1147793431 M * ptl_xid yeah, I paste just 2 lines =P 1147793447 M * ptl_xid Kernel: 2.6.12.5 1147793448 M * ptl_xid VS-API: 0x00020001 1147793448 M * ptl_xid util-vserver: 0.30.208; Sep 26 2005, 16:51:58 1147793448 M * derjohn ptl_xid, i mean cloumns ;) 1147793479 M * ptl_xid I have just 4 data columns, and 1 is the name 1147793513 M * ptl_xid is something wrong with my /proc/virtual ? 1147794251 M * daniel_hozac you just have really old versions :) 1147794356 M * brc_ ptl_xid: you are doing wrong it is unlimited 1147794358 M * brc_ -1 means unlimited 1147794400 M * brc_ create the file /etc/vservers/server_name/rlimits/rss.hard 1147794439 M * doener derjohn: sure, but unification includes setting IUNLINK AFAIK 1147795873 J * bonbons ~bonbons@83.222.38.81 1147796050 Q * Viper0482 Ping timeout: 480 seconds 1147796395 P * ptl_xid 1147796606 J * Viper0482 ~Viper0482@p54975AA2.dip.t-dialin.net 1147796785 N * otaku42 otaku42_away 1147797229 Q * bonbons oxygen.oftc.net xenon.oftc.net 1147797229 Q * alamar oxygen.oftc.net xenon.oftc.net 1147797229 Q * micah oxygen.oftc.net xenon.oftc.net 1147797229 Q * softi42 oxygen.oftc.net xenon.oftc.net 1147797229 Q * lonewolff oxygen.oftc.net xenon.oftc.net 1147797229 Q * Wonka oxygen.oftc.net xenon.oftc.net 1147797229 Q * blackfire oxygen.oftc.net xenon.oftc.net 1147797246 J * bonbons ~bonbons@83.222.38.81 1147797246 J * alamar ~alamar@sol.diddens.de 1147797246 J * micah ~micah@208.99.202.72 1147797246 J * softi42 ~softi@p549D61FE.dip.t-dialin.net 1147797246 J * lonewolff lonewolff@adleman.lonewolff.info 1147797246 J * Wonka produziert@chaos.in-kiel.de 1147797246 J * blackfire blackfire@dp70.internetdsl.tpnet.pl 1147797301 J * lonewolf1 lonewolff@adleman.lonewolff.info 1147797340 J * f_ ~f_@83-215-237-1.seek.stat.salzburg-online.at 1147797352 Q * f_ Quit: 1147797420 Q * lonewolff Ping timeout: 480 seconds 1147799113 J * Oli ~skycode@212.224.226.58 1147799208 M * derjohn My hist has CONFIG_NFS_V3=y , module nfs loaded, but mount says: NFSv3 not supported! ... ? 1147799219 M * derjohn (within the guest ...) 1147799546 J * click click@ti511110a080-5135.bb.online.no 1147799691 N * Bertl_zZ Bertl_oO 1147799745 M * daniel_hozac derjohn: i assume that it works from the host? did you give the guest the right ccaps? 1147799783 Q * lilalinux Read error: Connection reset by peer 1147799783 M * derjohn daniel_hozac: derjohn slaps himself. no caps set. do I need cap_sys_admin to mount? 1147799828 M * daniel_hozac well, you'll probably want the binary_mount ccap. 1147799832 M * daniel_hozac (http://linux-vserver.org/Caps+and+Flags) 1147800080 M * derjohn eh, so I have to write "0x00040000" into /etc/vservers/foo/ccapabilities ? 1147800101 M * daniel_hozac no, just binary_mount works fine ;) 1147800126 M * derjohn but really "ccapabilities" or "bcapabilities" 1147800153 M * daniel_hozac ccapabilities. 1147800342 J * shedi ~siggi@inferno.lhi.is 1147800380 J * lilalinux ~plasma@dslb-084-058-255-236.pools.arcor-ip.net 1147800582 Q * bubulak Ping timeout: 480 seconds 1147800761 Q * shedi Quit: Leaving 1147800793 J * bubulak ~bubulak@cicka.wnet.sk 1147800950 M * derjohn how can I check which capabilities a guest has? (in case i misspelled somehting etc) 1147801514 N * lonewolf1 lonewolff 1147802901 Q * Oli Quit: Oli 1147803246 J * shedi ~siggi@dsl-og-108-50.du.vortex.is 1147803515 Q * VooDooMaster Quit: Never put off till tomorrow, what you can do the day after tomorrow 1147803597 Q * eyck Quit: leaving 1147804062 Q * lilalinux Remote host closed the connection 1147804270 J * jpacheco ~justin@CPE00146c1608af-CM00159a655a08.cpe.net.cable.rogers.com 1147804276 M * jpacheco hey everyone :) 1147804294 M * jpacheco has anyone here gotten lvm + quota in a vhost working? 1147804324 M * jpacheco i got lvm working ok (no problems), but enabling quota support is proving to be difficult 1147804386 M * jpacheco i followed the following guides: http://linux-vserver.org/Standard+non-shared+quota, http://www.5dollarwhitebox.org/wiki/index.php/Howtos_Linux-Vserver_With_LVM_And_Quotas 1147804423 J * lilalinux ~plasma@dslb-084-058-255-236.pools.arcor-ip.net 1147804981 Q * Viper0482 Quit: bin raus, 1147805091 J * eyck eyck@ghost.anime.pl 1147805371 J * Oli ~skycode@212.224.226.58 1147805492 M * brc_ jpacheco: whichi kernel version are yuo using ? 1147805531 M * brc_ user/group disk quotas on a shared partition is still not supported by the 2.x branch 1147805695 M * derjohn daniel_hozac, at least i need ccapa secure_mount, but: "NFSv3 not supported!" (and yes, from the host it works) 1147805708 Q * eyck Ping timeout: 480 seconds 1147805829 J * eyck eyck@ghost.anime.pl 1147807645 J * NetAsh ~NetAsh@88.222.136.221 1147807958 P * NetAsh 1147808026 J * dearauj1 ~dearaujo@pixpat.austin.ibm.com 1147808208 M * derjohn is it normal behavior, that a guest is started in runlevel 3 by default? For debian runlevel 2 should be the way to go 1147809010 M * derjohn bye, back l8ter 1147809487 P * dearauj1 1147809634 J * ViRuS ~virus@86.34.56.115 1147809684 Q * ViRuS Quit: 1147809687 J * ViRuS ~virus@86.34.56.115 1147809692 Q * ViRuS Quit: 1147810616 Q * shedi Quit: Leaving 1147810661 Q * bonbons Quit: Leaving 1147810985 M * Loki|muh hmmm I have an odd issue - removing an lvm-volume gives me this error: Can't remove open logical volume "vroot2" 1147810996 M * Loki|muh but the volume isn't mounted any more 1147811062 M * Loki|muh and there is no vserver-guest running on this machine 1147813304 M * tokkee derjohn: $ grep initdefault /var/lib/vservers/vs-test/etc/inittab 1147813304 M * tokkee id:2:initdefault: 1147813458 M * tokkee Loki|muh: Which kernel are you running? I have a similar problem with unionfs... 1147813484 M * tokkee Loki|muh: However, I don't really think, they are related ;-) 1147813535 M * BenBen refoo # 1147813572 M * Loki|muh tokkee: 2.6.15.1-vs2.1.0.5 1147813583 M * Loki|muh i will try a newer one 1147813637 M * tokkee Hey BenBen ;-) 1147813667 M * tokkee Loki|muh: I have 2.6.14.3-vs2.0.1... 1147813690 M * tokkee BenBen: Who's that girl you where talking about in #vserver-linuxtag? 1147813759 M * tokkee Loki|muh: Hum... sounds like an LVM issue to me... :-/ 1147813989 M * Skram anyone around? 1147813990 Q * Hollow Read error: Connection reset by peer 1147813991 M * Skram well duh 1147813992 M * Skram okay 1147814010 A * tokkee is around ;-) 1147814048 J * Hollow ~hollow@home.xnull.de 1147814102 M * Skram all of a sudden it looks like all vpses are using half of what they used to use 1147814113 M * Skram half of the RSS 1147814164 M * Skram http://paste.linux-vserver.org/43 1147814169 M * Skram Will someone please look over that? 1147814243 A * doener goes looking up how awk works... 1147814325 A * tokkee will go to bed now ;-) 1147814331 M * Skram please, someone help 1147814370 A * tokkee is not able to give any usefull hints any more :-/ 1147814375 M * Skram ok 1147814384 M * Skram Bertl_oO: ? Hollow: ? 1147814405 M * Skram all of a sudden ram usage is like way lower 1147814408 M * Skram hercules / # free -m total used free shared buffers cached 1147814414 M * Skram Mem: 2026 841 1184 0 62 502 1147814421 M * Skram there used to be only like 100mb left 1147814542 M * doener did you change anything? was it really RSS or maybe just buffers/cache? 1147814560 M * Skram i don tthink i changed anything 1147814565 M * Skram and now free is showing differently 1147814706 M * doener the second line of free's output is more important. cache/buffers will be freed if a process starts using lots of memory. if that process ends, the memory will be free afterwards. the second line shows the used memory minus the values for buffers/cache and thus will not be affected by such a process (once it ended) 1147814796 M * Skram -/+ buffers/cache: 278 1748 1147814804 M * Skram im running 10+ vpses 1147814815 M * Skram and most have mail, apache, asterisk, and ftp 1147814817 M * Skram with users 1147814819 M * Skram how is that possible? 1147814834 M * doener unified? 1147814842 M * Skram havent tried that 1147814852 M * Skram what is it exactly? 1147814854 M * Skram *shrug* 1147814883 M * doener identical files are replaced by hardlinking a single copy of that file, lower disk and memory usage 1147814903 M * Skram i have never enabled that 1147815082 M * doener vserver with MySQL, Apache 2, SSH running takes about 13MB RAM here (according to free -m) 1147815183 M * Skram really? 1147815186 M * Skram weird. 1147815196 M * Skram i restarted most of the vpses and shit seems to be back to normal 1147815214 M * Skram sorry for the language- has been a crazy day 1147815250 M * doener well, I just started a few vservers and didn't do anything with them, so most allocated memory probably was still VSZ-only (zero-page) not RSS 1147815278 M * Skram hmm okay 1147815300 M * Skram will vserver auto-unify stuff? 1147815416 M * nebuchadnezzar Skram: no, you need to vserver vhashify for all 1147815421 M * Skram hmm ok 1147815523 J * shedi ~siggi@inferno.lhi.is 1147815922 Q * dna Quit: Verlassend 1147816365 Q * stefani Quit: ircII EPIC4-2.2 -- Are we there yet? 1147816546 M * brc_ Skram: how much ram do you have ? 1147816582 M * brc_ Skram: Try vmstat 1 1147816584 M * brc_ and see si / so 1147816778 M * daniel_hozac derjohn: and binary_mount ;) 1147816821 M * daniel_hozac derjohn: and i think 3 was the one most distros have in common as a "start everything but X" kind of level. 1147816826 M * Skram brc_: 1147816837 M * Skram procs -----------memory---------- ---swap-- -----io---- --system-- ----cpu---- r b swpd free buff cache si so bi bo in cs us sy id wa 0 0 111132 750204 203020 696084 1 1 2 3 1 1 6 8 85 1 0 0 111132 750080 203020 696084 0 0 0 0 1028 3548 0 0 100 0 1147816868 M * Skram si and so are 1 1147816874 M * Skram when i do "vmstat" 1147817056 Q * Oli Quit: Oli 1147817234 M * brc_ So if you feel slowdowns it is not memory related 1147817791 J * pzystorm schak@dslb-082-083-037-235.pools.arcor-ip.net 1147818113 J * s0undt3ch ~s0undt3ch@bl7-247-155.dsl.telepac.pt 1147818900 J * mire ~mire@200-167-222-85.COOL.ADSL.VLine.Verat.NET 1147818903 M * pzystorm is there a possibility (with a cap or so) to give the permission to mount something inside a guest? for example: a file with a loop-device which a put inside the guest 1147819267 M * derjohn daniel_hozac, # cat /etc/vservers/aps-qi-2-virtual/ccapabilities 1147819267 M * derjohn SECURE_MOUNT 1147819267 M * derjohn SECURE_REMOUNT 1147819267 M * derjohn BINARY_MOUNT 1147819267 M * derjohn ... I didnt work. no I mounted it from host and made a bind-mount into the guest (I am scared NFS goes down and stay 'unmountable' until i restart the guest) 1147819355 M * pzystorm 4008153777738 1147819373 M * pzystorm sorry ;o 1147819819 M * s0undt3ch hello ppl 1147819829 M * s0undt3ch is it possible to rename a vserver? 1147819836 M * Skram while running? no 1147819846 Q * matti Ping timeout: 480 seconds 1147819850 M * s0undt3ch Skram: and while not running, how? 1147819872 M * Skram change the /vserver/name, /etc/vserver/name 1147819885 M * s0undt3ch Skram: that's it? 1147819891 M * Skram and edit uts/nodename as well as the symbolic links in /etc/vservers/name/ 1147819914 M * s0undt3ch where's uts/nodename? 1147819945 M * derjohn s0undt3ch, it only there is you set a nodename (== hostname within a guest) 1147819968 M * Skram /etc/vservers/uts/nodename 1147819974 M * Skram how do you run vservers and not know? 1147819976 M * s0undt3ch oh, k thanks 1147819980 M * Skram just crious 1147819992 M * s0undt3ch Skram: I run them for about a week or two :) 1147820001 M * Skram /// 1147820002 M * Skram ok 1147820009 M * derjohn s0undt3ch, but be aware that there are symlinks within /etc/vservers/oldname/vdir... which points to /var/lib/vservers/oldname 1147820010 M * s0undt3ch just changed to vserver 1147820028 M * Skram ok 1147820031 M * derjohn s0undt3ch, if you want to change the dir name, dont forget the symlinks 1147820073 M * s0undt3ch well, I think I'll leave it as is for now, too much unkown problems(to me) may occurr 1147820087 M * s0undt3ch derjohn: you have a vserver taking care of ldap for you right? 1147820139 M * derjohn s0undt3ch, yes and no. i have a all-in-one web-host-o-matic guest, with ISPman and LDAP backend ... 1147820202 J * Dr4g ~Dr4g@82-40-202-184.stb.ubr06.uddi.blueyonder.co.uk 1147820215 M * s0undt3ch derjohn: I'll be trying to make a vserver take care of all auth/hosts/nis stuff for all other guests and also if possible to the vserver 1147820256 M * s0undt3ch seen any info on that subject? 1147820290 M * derjohn s0undt3ch, shouldnt be a problem. it has nothing to do with vserver in general , but stand auth stuff. 1147820324 M * derjohn s0undt3ch, well, youl'll need the migration scripts vom PDAL, openldap package, and libnss-ldap (or libpam-ldap) 1147820356 M * derjohn s0undt3ch, then edit /etc/nsswitch.conf, replace compat with "compat ldap" and reboot (the guests) 1147820365 M * derjohn s0undt3ch, But: YMMV ! 1147820374 M * s0undt3ch yep untill there I know 1147820378 M * s0undt3ch YMMV? 1147820382 M * derjohn s0undt3ch, infos? openldap.org 1147820393 M * derjohn YMMV = Your Milage may vary. 1147820400 M * derjohn ah 1147820401 M * s0undt3ch k ;) 1147820428 M * derjohn s0undt3ch, if you did that, type "getent passwd" , then you see the 'virtual' (NSS-ed) passwd file 1147820442 M * derjohn s0undt3ch, if this is cool, then you won ;) 1147820460 M * s0undt3ch I think my main preocupation is the fact that i'll try to make it also handle the root vserver 1147820471 M * s0undt3ch derjohn: still didn't do it 1147820476 M * s0undt3ch I'm about to ;) 1147820493 M * derjohn thhere are several tutorials out on the web if you google for one of the specific terms i mentioned. 1147820535 M * derjohn s0undt3ch, the root-server simply uses als ldap host in /etc/libnns.conf 1147820541 M * derjohn libnss-ldap.conf 1147820548 M * s0undt3ch k, thanks 1147820563 M * s0undt3ch I'll bug you if I get in trouble ;) 1147820573 M * derjohn s0undt3ch, OMG ;) 1147820579 M * derjohn btw 1147820600 M * derjohn s0undt3ch, #ldap on freenode is the place to go for ldap problems 1147820620 M * derjohn s0undt3ch, and maybe 'luma' is cool for editing the ldap tree 1147820642 M * s0undt3ch he he he, I know, but you aparently done this :) don't worry I won't bug you too much 1147820660 M * s0undt3ch k, i'll also google for luma 1147820674 M * derjohn s0undt3ch, feel free to ask. 1147820682 M * s0undt3ch thanks 1147820686 M * derjohn luma is an kde based graphical ldap-editot-foo 1147820689 M * derjohn k 1147820695 M * s0undt3ch oh, kde :| 1147820702 M * s0undt3ch I don't run kde 1147820711 M * s0undt3ch nor will I ever run 1147820728 M * s0undt3ch been there and simply didn't like it 1147820740 M * derjohn s0undt3ch, there Java based thingy out (Jexplre or such ...) and GQ or so for gnome 1147820759 M * derjohn s0undt3ch, but you can use ldapsearch/modify from shell, too 1147820770 M * s0undt3ch I saw one yesterday, I think it's called web2ldap 1147820779 M * derjohn s0undt3K one day you will ;) 1147820796 M * s0undt3ch well, who knows ;) 1147820806 M * derjohn yeah, thre are php based ones out, too 1147820816 M * s0undt3ch this one is python based 1147820850 M * derjohn YMMV ;) 1147820863 M * s0undt3ch yep 1147820874 M * s0undt3ch well, getting my hands dirty 1147820875 M * Loki|muh tokkee: wasn't a lvm-issue, I'm using now 2.6.16.16-vs2.1.1-rc20 and it works :) 1147820926 Q * Wenix Ping timeout: 480 seconds 1147821089 J * sventon ~aaa@c-7bed70d5.021-60-6c756e10.cust.bredbandsbolaget.se 1147821091 P * sventon 1147821315 J * Aiken ~james@tooax6-138.dialup.optusnet.com.au 1147821453 Q * michal` Ping timeout: 480 seconds 1147821694 J * michal` ~michal@www.rsbac.org 1147822031 Q * harry Ping timeout: 480 seconds 1147822450 M * brc_ hey 1147822453 M * brc_ from the vserver paper: 1147822453 M * brc_ The only drawback is that without additional measures, a malicious context would be able to deliberately or accidentally destroy or modify such shared files, which in turn would harm the other contexts. 1147822460 M * brc_ - 1147822468 M * brc_ 6.0. Unification 1147822499 M * brc_ If i just use vunify or vhashify won't it be safe/ :) 1147822766 M * brc_ Weird 1147822766 M * brc_ servidor:~/vserver/util/util-vserver-0.30.210# /usr/lib/util-vserver/vhashify demo_vhcs 1147822766 M * brc_ Segmentation fault 1147822917 M * doener brc_: you missed the "without additional measures" part ;) of course those are taken with Linux-VServer 1147822956 M * doener I don't think you should use vhashify directly... after preparing the vserver for hashification, use "vserver foo hashify" 1147822983 M * doener http://linux-vserver.org/alpha+util-vserver -- take a look at the vhashify section on that page 1147823288 M * brc_ That's what i am reading :) 1147823297 M * brc_ This unification stuff seems really nice 1147823307 M * brc_ I am excited to see how much memory will be freed 1147823311 M * brc_ and how faster the machine will be 1147823508 M * doener memory might be quite a bit if there are big programs/libraries in use that are running in lots of vserver... faster? hm, maybe less cache misses if there are many context switches between these processes in different vservers, but the memory will probably have a bigger effect if it allows better/more disk caching 1147823565 M * brc_ I have servers splitted into two partitions. If i knew i was going to unify it shouldn't have done it 1147823605 M * brc_ i am getting segfault. weird 1147823617 M * doener with "vserver foo hashify" as well? 1147823621 M * doener hum... 1147823644 M * brc_ Initializing exclude-list for /fs5/vservers/demo_vhcs (demo_vhcs) 1147823644 M * brc_ Starting to traverse directories... 1147823644 M * brc_ Segmentation fault 1147823665 M * brc_ it has somethign to do with the partition in use.. any hints ? 1147823874 M * doener no, I have no idea about how vhashify works