1193616055 J * friendly12345 ~friendly@ppp59-167-134-217.lns3.mel6.internode.on.net 1193618317 Q * friendly12345 Quit: Leaving. 1193619293 J * friendly12345 ~friendly@ppp59-167-134-217.lns3.mel6.internode.on.net 1193619805 Q * derjohn_mobil Ping timeout: 480 seconds 1193620201 Q * mire Ping timeout: 480 seconds 1193620501 Q * meandtheshel1 Quit: Leaving. 1193622085 Q * hparker Quit: peer reset by connection 1193622199 Q * faheem_ Ping timeout: 480 seconds 1193623847 J * hparker ~hparker@linux.homershut.net 1193624114 Q * Aiken Quit: Leaving 1193625572 J * Aiken ~james@ppp121-45-228-65.lns1.bne4.internode.on.net 1193626024 J * RedfernS ~chatzilla@trts01.totalrisc.com.au 1193626994 M * RedfernS Greetings vservers, i have a silly question. The file provided is a diff file. This is my first time dealing with this file type. What is the command and switches that you use to apply it to a custom kernel that you are building? 1193627061 M * daniel_hozac http://linux-vserver.org/Installation_on_Linux_2.6 1193627210 J * fatgoose ~samuel@206-248-128-87.dsl.teksavvy.com 1193627231 M * RedfernS Thanks daniel_hozac 1193629381 Q * fatgoose Quit: fatgoose 1193633038 J * arachnist arachnist@088156184167.who.vectranet.pl 1193634096 J * balbir ~balbir@59.178.36.100 1193635035 Q * balbir Ping timeout: 480 seconds 1193635725 J * balbir ~balbir@59.178.36.100 1193636266 Q * balbir Ping timeout: 480 seconds 1193636997 J * balbir ~balbir@59.178.41.32 1193637859 Q * balbir Ping timeout: 480 seconds 1193638396 Q * hparker Remote host closed the connection 1193638588 J * balbir ~balbir@59.178.62.93 1193639052 Q * balbir Read error: Operation timed out 1193640147 J * mire ~mire@118-168-222-85.adsl.verat.net 1193640274 Q * RedfernS Quit: ChatZilla 0.9.78.1 [Firefox 2.0.0.8/2007100816] 1193640626 N * ensc Guest280 1193640636 J * ensc ~irc-ensc@p54B4EADB.dip.t-dialin.net 1193640745 Q * Guest280 Ping timeout: 480 seconds 1193640845 J * balbir ~balbir@59.178.57.222 1193641275 J * DuckMaster ~Duck@81.57.39.234 1193641275 Q * duckx Read error: Connection reset by peer 1193641507 A * Supaplex checks the wiki, how to allow usb device access for a specific device humm... 1193642421 J * besonen_mobile ~besonen_m@71-220-228-70.eugn.qwest.net 1193643467 J * dna ~dna@175-199-dsl.kielnet.net 1193643523 Q * dna 1193644188 J * JonB ~NoSuchUse@kg1-98.kollegiegaarden.dk 1193644541 M * harry someone here? 1193644905 J * pmenier ~pme@LNeuilly-152-22-72-5.w193-251.abo.wanadoo.fr 1193644952 Q * larsivi Ping timeout: 480 seconds 1193645105 M * JonB yes 1193645145 M * harry why is there a 16 ip address limit in a vserver? 1193645193 M * JonB harry: i do not fully understand your question 1193645219 M * harry i have a vserver guest 1193645227 M * JonB yes? 1193645228 M * harry i want to give it 21 ip adresses 1193645257 M * JonB on how many network devices? 1193645258 M * harry (it's a webserver, serving about 100 sites, 20 of which are https) 1193645261 M * harry 1 1193645272 M * JonB okay 1193645319 M * JonB does regular linux have a limit like this? 1193645367 M * harry no 1193645407 Q * mire Read error: Operation timed out 1193645409 Q * balbir Read error: Operation timed out 1193645410 M * JonB what happens if you make ethernet device aliases? 1193645428 M * JonB can you then have more ip addresses in a vserver guest? 1193645450 M * JonB do you only have this problem for https? or also ssh? 1193645567 M * harry ? 1193645584 M * JonB i am trying to think of other work arrounds 1193645594 M * harry the vserver won't start if i have more than 16 directories in the /etc/vservers//interfaces dr 1193645597 M * harry dir 1193645608 M * harry you can't define aliases... vserver uses network isolation 1193645616 M * harry but with a 16 ip address limit 1193645648 M * JonB like using iptables to redirect the packets to those ip addresses to the server 1193645660 M * JonB but i think it will then reply with the other address 1193645675 M * JonB maybe the problem is in the vserver-utils package 1193645694 M * JonB to my understanding the vserver-utils package is made by daniel_hozac 1193645701 M * JonB and the kernel mostly by Bertl_zZ 1193645723 J * derjohn_mobil ~aj@e180203117.adsl.alicedsl.de 1193645742 J * DoberMann ~james@cap31-6-88-180-72-76.fbx.proxad.net 1193645752 J * DavidS ~david@p54812599.dip0.t-ipconnect.de 1193645784 Q * DoberMann 1193645804 M * JonB harry: i do not feel i can help you (more) 1193645863 M * harry yeah 1193645864 M * harry i know :) 1193645868 M * harry tnx anyways :) 1193645880 A * harry looks in the souce... if only i knew where to look :) 1193645884 M * harry but... on my way ;) 1193645921 J * DoberMann ~james@cap31-6-88-180-72-76.fbx.proxad.net 1193645929 M * JonB ok 1193646129 J * balbir ~balbir@59.178.36.128 1193646455 P * DoberMann 1193647329 Q * balbir Ping timeout: 480 seconds 1193647374 Q * JonB Quit: This computer has gone to sleep 1193647992 J * balbir ~balbir@59.178.40.247 1193648458 J * JonB ~NoSuchUse@kg1-98.kollegiegaarden.dk 1193648939 J * gebura ~gebura@77.192.186.197 1193649286 J * torsti76 ~torsti76@gate.iwm-kmrc.de 1193649317 M * torsti76 hi all 1193649340 M * torsti76 daniel_hozac or bertl, is one of you there? 1193649364 M * torsti76 i have a serious problem with one of our vservers on gentoo: 1193649426 N * Bertl_zZ Bertl 1193649432 M * Bertl morning folks! 1193649439 M * torsti76 morning bertl 1193649471 M * fb_ morning Bertl 1193649473 M * torsti76 bertl, may i consult you with some questions? 1193649516 M * torsti76 i have a vserver on a gentoo host (latest kernel, latest utils), which ran until recently, when i updated the world inside the vserver 1193649553 M * Bertl Hollow is the gentoo expert, and it sounds very gentoo specific to me :) 1193649563 M * torsti76 this hasn't been done for quite a while, so just to make sure all services (in particular apache and ssh) are running stable and well, i tried to restart the vserver 1193649604 M * torsti76 and now, even with NO service in any of the runlevels, the vserver restarts every 3 seconds or so 1193649628 M * Bertl even with, or exactly because? 1193649643 M * torsti76 so you can just enter it and after a short while it spits you out to the host with "killed" 1193649653 M * torsti76 even with 1193649663 M * torsti76 first trial was with everything in place 1193649669 M * Bertl if you are using the sysv init style equivalent for gentoo (called 'gentoo') 1193649677 M * torsti76 the "older" vservers on the same host run just fine 1193649694 M * torsti76 until now i used "plain" 1193649696 M * Bertl then, IIRC, you have to have at least one service enabled to keep the guest alive 1193649705 M * torsti76 i c 1193649709 M * Bertl if you still use plain, then you should be fine 1193649713 M * torsti76 will try, just a second 1193649730 M * weasel any news on 2.6.23? 1193649730 M * Bertl plain will start an init process inside the guest, which should run even without any service 1193649731 M * torsti76 hmm, that cancels that out 1193649751 M * Bertl weasel: nope, unfortunately I had other things to do 1193649753 M * torsti76 may it be a problem with glibc-2.6.1? 1193649757 M * weasel too bad 1193649784 M * Bertl weasel: any deadline you are faced with? 1193649830 M * Hollow torsti76: which baselayout version do you have now? 1193649844 M * torsti76 i even tried the following: chrooted into the vserver, ran some installations, tried some processes, etc. - init unfortunately is not that easy to try out ;o) 1193649853 M * torsti76 hollow: just a sec 1193649879 M * torsti76 1.12.6 1193649897 M * Hollow so, i assume you have baselayout-vserver? 1193649907 M * torsti76 you assume right ;o) 1193649934 M * Hollow i'd suggest to update to baselayout-2.0.0_rc5 then 1193649944 M * Hollow it is vserver compatible out of the box 1193649954 M * torsti76 is it already in portage? 1193649960 M * Hollow yes, but masked 1193649964 M * torsti76 i'll give it a shot 1193649972 M * torsti76 will report back here 1193649995 M * Hollow torsti76: before you update... 1193650003 M * Hollow you have to switch the portage profile too 1193650024 M * Hollow i.e. do not use the vserver profiles anymore, just use default-linux/x86/2007.0 1193650046 M * torsti76 ok 1193650064 M * Hollow and it is also a good idea to remove all the initscript that symlink to dummy before updating 1193650066 M * torsti76 but doesn't that pull the "usual" baselayout, too? 1193650073 M * Hollow yes, that's the ide 1193650074 M * Hollow a 1193650082 M * torsti76 ok ;o) 1193650093 M * Hollow just make sure it is 2.0.0_rc5 1193650247 M * weasel Bertl: not really, no 1193650568 P * friendly12345 1193650817 M * torsti76 hollow: still one question 1193650852 M * torsti76 before upgrading to baselayout, should i unmerge baselayout-vserver or will that render the server unusable? 1193650887 M * Hollow i would suggest to install the newer baselayout first, although i have done the other way round too .. if you know what you are doing both work ;) 1193650925 M * torsti76 then i assume i don't know, what i'm doing, albeit being with gentoo for about 6 years ;o) 1193650947 M * Hollow the really important thing is to remove the symlinks in /etc/init.d that point to dummy 1193650958 M * Hollow etc-update isn't smart enough to replace symlinks 1193650991 M * torsti76 yes, did that already 1193651028 M * torsti76 anyway - nice to finally have a vserver-aware baselayout 1193651034 M * Hollow yes, indeed 1193651044 M * torsti76 that saves lot of trouble when migrating, too 1193651068 M * Hollow it was hard work to make it work out of the box ;) 1193651079 M * Hollow but it works really good now on my boxes 1193651081 M * torsti76 and, btw., that was the whole idea before updating the server -> move a copy to a real machine 1193651092 M * torsti76 yes, i would guess so 1193651163 M * Hollow you might also want to run /usr/lib/util-vserver/distributions/gentoo/initpost /etc/ /usr/lib/util-vserver/util-vserver-vars 1193651171 M * Hollow after you have upgraded and stopped the guest 1193651180 M * Hollow this will do some cosmetic cleanup 1193651191 M * Hollow but should not be needed for plain ini style 1193651298 J * larsivi ~larsivi@85.221.53.194 1193651306 M * gebura hi 1193651353 J * DoberMann ~james@cap31-6-88-180-72-76.fbx.proxad.net 1193651363 Q * DoberMann 1193651377 J * DoberMann ~james@cap31-6-88-180-72-76.fbx.proxad.net 1193651652 M * Bertl hey gebura! wb larsivi! 1193651872 Q * larsivi Remote host closed the connection 1193651908 J * larsivi ~larsivi@85.221.53.194 1193652024 Q * DoberMann Quit: brb, avec le plugin-qui-va-bien-pour-supporter-les-utf8eurs :p 1193652039 J * DoberMann ~james@cap31-6-88-180-72-76.fbx.proxad.net 1193652366 J * zLinux ~zLinux@88.213.32.223 1193652509 Q * larsivi Remote host closed the connection 1193652527 J * larsivi ~larsivi@85.221.53.194 1193652684 P * DoberMann 1193652877 M * harry Bertl: !!!!! 1193652879 M * harry is you there? 1193653442 M * Supaplex hi harry 1193653553 M * Bertl harry: mesa is there :) 1193653610 M * Supaplex the wiki says vserver is some 6800 lines of patch? 1193653685 M * Supaplex I've got a few people hooked on it now :) 1193653753 M * Supaplex i have some kind of intermittent routing issue fixed now (I hope). when i'm up again, I'll look into drbd 1193653783 M * harry Bert? 1193653803 M * harry Bertl: is there a reason to have only 16 ip addresses in a guest? 1193653810 M * harry why that limit? 1193653816 M * Bertl you mean, for older kernels, yes? 1193653827 M * harry if you call 2.6.22.10 older ;) 1193653845 M * Bertl no, but 2.2.x, 2.3.x doesn't have that limit anymore 1193653847 M * harry and util-vserver 0.30.213 1193653872 M * harry my server won't start if i add more than 16 ips 1193653895 M * Bertl vs2.3.x and util-vserver 0.30.214 should be fine# 1193653897 M * mjt 2.6.22 IS older than 2.6.32... so that's an OLDER kernel indeed (not OLD but older than "current" anyway) ;) 1193653902 M * mjt s/32/23/ 1193653931 M * harry 2.3 is development... it's a production server 1193653935 A * harry upgrades the tools 1193653938 M * harry will that help? 1193653947 M * Bertl production servers should be fine with 16 ips :) 1193653953 M * harry no they are not 1193653962 M * harry we need a lot of https websites 1193653971 M * harry different domains ==> different ips 1193653973 M * Bertl then you have to patch ... 1193653975 M * harry 16 is NOT enough 1193653980 M * harry :) 1193653992 J * lidi20 ~lidi20@ANantes-257-1-96-67.w90-25.abo.wanadoo.fr 1193653992 M * Bertl there are patches floating around for 64 and more IPs 1193654006 M * Bertl note, that you will have some performance overhead 1193654010 M * Bertl welcome lidi20! 1193654069 M * mjt there's a patch queued(?) for 2.6.24 to add an "ifgroup" framework, to group interfaces (to see them, say, in iptables as a single unit). Maybe something similar may be used for vserver too? 1193654084 M * mjt for vserver IP addresses, that is 1193654086 M * harry real 0m10.897s 1193654086 M * harry user 1m12.690s 1193654086 M * harry sys 0m44.410s 1193654092 M * harry utilvserver compile :) 1193654118 M * mjt harry: that seems broken 1193654123 Q * lidi20 1193654123 M * mjt real time can' 1193654141 M * harry mjt: ahm... why broken? 1193654145 M * mjt real time can't be less than sum of user and sys 1193654145 M * harry that's perfectly possible 1193654151 M * harry why not? 1193654155 M * Bertl mjt: sure, multiple CPUs 1193654157 M * harry (i have 8 cores) 1193654170 M * mjt err.. I stand corrected. 1193654188 M * mjt never noticied it behaves like that on multi-cpu 1193654205 M * harry mjt: you have to make -jX for more threads to be used 1193654212 M * mjt sure 1193654239 M * mjt for large compiles i usually use multiple machines (32..40 cores total) 1193654443 M * harry so Bertl , it's only in 2.3 that you can have more than 16 ip addresses? 1193654450 M * harry why and where is that limit set? 1193654511 M * Bertl no, as I said above, you can patch older version to have more than 16 ips 1193654526 M * Bertl but the 16ip limit was choosen as appropriate upper limit 1193654529 M * harry where is that patch? 1193654542 M * Bertl use google, off for lunch now ... back later 1193654544 M * harry it's a kernel limit? 1193654547 M * harry cya 1193654549 N * Bertl Bertl_oO 1193655250 N * esa eSa| 1193655577 M * brc does vserver have any kind of live migration? 1193655732 Q * eSa| Ping timeout: 480 seconds 1193655819 J * lilalinux ~plasma@80.69.41.3 1193655832 M * torsti76 hollow: i now installed 2.0.0_rc5 - had a meeting afterwards, that's why i'm getting back to you only now 1193655855 M * torsti76 it works - up to a certain point ;o) 1193655878 M * torsti76 it doesn't seem to recognize default as the desired runlevel 1193655900 M * torsti76 instead, eselect rc show gives the following: 1193655916 M * torsti76 webex / # eselect rc show 1193655916 M * torsti76 Status of init scripts in runlevel sysinit 1193655963 M * torsti76 all init scripts in "boot" and "default" seem have status stopped 1193655967 M * AStorm torsti76: you forgot to set init style to GEntoo 1193656014 M * Hollow torsti76: i'm not sure the eselect module supports baselayout2 1193656018 M * Hollow try rc-status 1193656042 M * torsti76 that couldn't be found 1193656049 M * Hollow uhm 1193656056 M * torsti76 and i think i found the reason for thze weird behavior 1193656092 M * torsti76 # /sbin/runscript doesn't exist either 1193656123 M * torsti76 is it part of the baselayout? 1193656150 M * AStorm yep 1193656171 M * torsti76 i'm just reinstallung the baselayout 1193656192 M * torsti76 seems, some things got washed away when i unmerged baselayout-vserver 1193656202 M * AStorm indeed 1193656208 M * AStorm baselayout-vserver = old 1193656229 M * torsti76 anyway, init-style gentoo shouldn't be able to do the trick 1193656236 Q * derjohn_mobil Ping timeout: 480 seconds 1193656243 M * torsti76 plain is just fine 1193656361 M * torsti76 if i set init/style to gentoo, i get an error: env: /lib/rcscripts/sh/init-vserver.sh: No such file or directory 1193656381 M * AStorm ahm yes 1193656391 M * AStorm it has some evil hacks for baselayout 1 1193656397 M * AStorm baselayout 2 shouldn't need that 1193656419 M * torsti76 i'll look at the inittab inside the vserver 1193656432 M * torsti76 the problem with the default runlevel should be there 1193656482 M * torsti76 hmm, it simply CAN'T work ;o) 1193656488 M * torsti76 the inittab is missing, too 1193656505 M * torsti76 what package does it belong to, usually? 1193656516 Q * JonB Quit: This computer has gone to sleep 1193656652 M * AStorm baselayout 1193656657 M * AStorm or sysvinit 1193656675 M * torsti76 sysvinit 1193656684 M * torsti76 already reinstalled it 1193656688 M * torsti76 now lets see 1193656731 M * torsti76 didn't help 1193656735 M * torsti76 hmm, lunchtime ;o) 1193657240 J * esa ~esa@ip-87-238-2-45.adsl.cheapnet.it 1193657270 N * Bertl_oO Bertl 1193657273 M * Bertl back now 1193657391 N * pmenier patdelephant 1193657399 N * patdelephant patpat 1193657424 N * patpat pmenier 1193658191 Q * balbir Ping timeout: 480 seconds 1193658559 Q * Aiken Quit: Leaving 1193658906 Q * DavidS Quit: Leaving. 1193658928 J * balbir ~balbir@59.178.58.20 1193659059 M * torsti76 i'm back 1193659082 N * pmenier pmenier_off 1193659106 M * torsti76 and back to my latest problem - how to convince init/style plain to accept the gentoo runlevels with baselayout2 1193659209 M * torsti76 any ideas? 1193659271 M * AStorm I convinced gentoo style to do so 1193659277 M * AStorm with minor patching 1193659304 M * torsti76 how did you do it? 1193659330 M * AStorm you have to copy some files 1193659469 Q * balbir Ping timeout: 480 seconds 1193659513 M * Bertl AFAIK, baselayout2 is supposed to fully support 'gentoo' init style 1193659547 M * Bertl and 'plain' init style should work out of the box with any distro, because it uses the distro's init 1193659547 M * torsti76 bertl, maybe, but it throws an error for me 1193659566 M * Bertl what kind of error` 1193659570 M * Bertl s`/? 1193659583 M * AStorm torsti76: you didn't install the vserver with gentoo init style 1193659586 M * torsti76 if i use init-style gentoo i get: 1193659593 M * AStorm so the files haven't been copied :> 1193659602 M * torsti76 env: /lib/rcscripts/sh/init-vserver.sh: No such file or directory 1193659626 M * torsti76 astorm, i understand 1193659639 M * Bertl yep, those seem to be missing, there is a post install script you can executed, IIRC 1193659655 M * torsti76 aaah, wait a sec 1193659659 M * Bertl s/executed/execute/ 1193659748 M * torsti76 do you mean the 'initpost' script? 1193659762 M * Bertl yep 1193659763 M * torsti76 i first have to locate it ;o) 1193659960 M * torsti76 bertl, that did it! 1193659984 M * torsti76 now everything comes up just fine with plain 1193660000 M * torsti76 gentoo still fails, but that shouldn't be my concern ;o) 1193660024 M * AStorm warning: that breaks the system 1193660033 M * AStorm because it replaces halt and reboot scripts 1193660068 M * torsti76 what do you mean? the script i just ran? 1193660087 M * Bertl torsti76: check again with Hollow, gentoo is supposed to work fine too 1193660099 J * balbir ~balbir@59.178.60.94 1193660225 M * Hollow torsti76: run /usr/lib/util-vserver/distributions/gentoo/initpost /etc/ /usr/lib/util-vserver/util-vserver-vars 1193660231 M * Hollow on the host 1193660242 M * Hollow this will copy the necessary files for gentoo init style 1193660287 M * torsti76 i did that just a minute ago 1193660302 M * torsti76 and still get env: /lib/rcscripts/sh/init-vserver.sh: No such file or directory 1193660313 M * Hollow did you set gentoo init style before? 1193660324 M * torsti76 nope, afterwards 1193660327 M * torsti76 just a sec 1193660473 M * torsti76 looks goooood 1193660480 M * Hollow yep :) 1193660511 M * torsti76 thank you all very much - i appreciate the fast help! 1193660553 M * Hollow remember to use the stages at http://people.linux-vserver.org/~hollow/ for new guests .. they already contain baselayout2 1193660613 M * Bertl okay, off for now .. back later ... 1193660619 N * Bertl Bertl_oO 1193660621 M * torsti76 hmm, IF we go on using vserver here - my boss got very attracted to xen ... 1193660662 M * Bertl_oO torsti76: well, if he likes to use fewer guests and buy more expensive hardware .. then Xen is probably a good choice too :) 1193660733 M * torsti76 i'll try to communicate ;o) 1193660751 M * torsti76 he seemed to be an intelligent species so far ;o) 1193660839 Q * balbir Ping timeout: 480 seconds 1193661059 J * derjohn ~derjohn@80.69.42.51 1193661444 M * AStorm Hollow: do these stages work as standalone systems too? (not just on vserver) 1193661511 J * balbir ~balbir@59.178.35.211 1193661563 M * Hollow AStorm: i think so .. but i didn't try yet :) 1193662503 J * lidi20 ~lidi20@ANantes-257-1-96-67.w90-25.abo.wanadoo.fr 1193662620 Q * lidi20 1193662769 J * derjohn_mobil ~aj@80.69.41.3 1193662909 J * virtuoso_ ~s0t0na@ppp91-122-160-203.pppoe.avangard-dsl.ru 1193663047 Q * balbir Read error: Operation timed out 1193663245 J * julien ~julien@crealab.info 1193663264 M * julien hello 1193663317 Q * virtuoso Ping timeout: 480 seconds 1193663330 Q * bzed Quit: leaving 1193663335 J * bzed ~bzed@devel.recluse.de 1193663521 Q * bzed 1193663530 J * bzed ~bzed@devel.recluse.de 1193663790 J * Julius ~julius@p57B26DFE.dip.t-dialin.net 1193663896 Q * pmenier_off Read error: Connection reset by peer 1193663907 J * pmenier_off ~pme@LNeuilly-152-22-72-5.w193-251.abo.wanadoo.fr 1193664020 J * balbir ~balbir@59.178.62.138 1193664114 J * Pazzo ~ugelt@reserved-225136.rol.raiffeisen.net 1193664430 J * JonB ~NoSuchUse@kg1-98.kollegiegaarden.dk 1193664445 J * ema ~ema@rtfm.galliera.it 1193665007 P * torsti76 1193665216 Q * derjohn_mobil Ping timeout: 480 seconds 1193665433 M * dowdle Good morning/evening. 1193665441 Q * gebura Quit: Quitte 1193665628 J * gebura ~gebura@77.192.186.197 1193665960 J * mire ~mire@118-168-222-85.adsl.verat.net 1193666076 Q * bzed Quit: leaving 1193666133 J * bzed ~bzed@devel.recluse.de 1193666231 M * AStorm dowdle: you mean generic timezone greetings? (because here it's neither morning nor evening :> ) 1193666261 J * toidinamai ~frank@svenska.toidinamai.de 1193666281 M * toidinamai Hi. 1193666296 M * gebura hi 1193666298 M * toidinamai Is there a way to go from one vserver to another without using SSH? 1193666311 M * daniel_hozac you mean, to break out of the vserver to get into another? 1193666317 M * daniel_hozac no. 1193666323 M * dowdle AStorm: It is to morning "here". 1193666360 M * mjt toidinamai: you mean `vserver $foo enter' from host? 1193666404 M * AStorm toidinamai: there might be if you expose certain devices 1193666413 M * AStorm but that's inherently unsafe and disabled by default 1193666437 M * toidinamai Ok, thanks. 1193666450 M * toidinamai SSH should support listening on unix domain sockets. 1193666470 M * AStorm why should it? Isn't localhost TCP good enough? 1193666485 M * toidinamai I though there is no localhost in vservers. 1193666487 J * derjohn_mobil ~aj@80.69.41.3 1193666496 M * AStorm toidinamai: there is if you add it 1193666501 M * AStorm and it's even shared between them 1193666515 M * toidinamai So no two vservers can listen on localhost:22? 1193666537 M * daniel_hozac this depends on your version. 1193666539 M * AStorm yes 1193666545 M * daniel_hozac 2.3 has a private 127.0.0.0/8 per guest. 1193666546 M * AStorm it's version-dependent 1193666552 M * toidinamai I use 2.2+grsec. 1193666554 M * AStorm daniel_hozac: is that optional? :> 1193666555 M * mjt there are whole 16777216 addresses on a loopback interface... 1193666591 M * daniel_hozac AStorm: yes. 1193666606 M * AStorm good. Any other way interfaces can be shared? 1193666626 M * mjt interfaces are NOT shared. IP addresses are 1193666633 M * daniel_hozac why would you want to? 1193666637 Q * derjohn Read error: Connection reset by peer 1193666645 M * daniel_hozac you don't want guests to interfere with eachother, do you? 1193666676 M * AStorm it may have some uses 1193666689 M * AStorm though not too many :> 1193666706 M * AStorm hypothetical question 1193666818 Q * bzed Quit: leaving 1193666824 J * bzed ~bzed@devel.recluse.de 1193666837 M * daniel_hozac just assign the same address to two guests... 1193666851 M * AStorm mhm 1193666861 M * toidinamai Is there a reference documentation for all files in /etc/vserver? 1193666873 M * daniel_hozac like the great flower page? 1193666877 M * AStorm toidinamai: great flower page 1193666881 M * AStorm daniel_hozac: :-) 1193666913 M * toidinamai Thanks! 1193666934 M * toidinamai I always forget about that when I haven't used vserver for a few months. 1193666966 M * toidinamai It should be linked from the Wiki main page. 1193666995 M * AStorm indeed 1193666996 J * derjohn ~derjohn@dslb-084-059-002-074.pools.arcor-ip.net 1193666998 M * daniel_hozac http://linux-vserver.org/util-vserver:Documentation 1193667034 M * toidinamai What about http://linux-vserver.org/Documentation ? 1193667041 J * lilalinux_ nas@dslb-084-059-002-074.pools.arcor-ip.net 1193667050 M * daniel_hozac that page is linked from there. 1193667068 M * toidinamai Ok. And /Documentation is linked from the main page. 1193667075 M * daniel_hozac it's in the toolbar on the left. 1193667095 M * AStorm yeah 1193667097 M * AStorm on the bottom 1193667103 M * AStorm ;> 1193667126 Q * larsivi Ping timeout: 480 seconds 1193667130 M * toidinamai Maybe the great flower page should be linked right next to the FAQ. 1193667136 M * toidinamai I think it's just as important. 1193667162 M * daniel_hozac util-vserver is just one userspace implementation. 1193667180 M * toidinamai ... 1193667194 M * AStorm so there are others? links please :> 1193667211 M * daniel_hozac http://svn.linux-vserver.org/projects/vcd 1193667305 M * AStorm eh, that, I hoped for something more 1193667482 Q * lilalinux Ping timeout: 480 seconds 1193668251 Q * FireEgl Read error: Connection reset by peer 1193668306 J * CWC CWC@89-215-37-177.2073053861.ddns-lan.pl.ekk.bg 1193668392 Q * CWC 1193669100 Q * balbir Ping timeout: 480 seconds 1193669772 J * balbir ~balbir@59.178.39.213 1193669792 J * Yvo yvonne@vpn229.rz.uni-mannheim.de 1193670184 Q * JonB Quit: This computer has gone to sleep 1193670455 P * Yvo 1193670524 J * tanjix tanjix@77.37.25.239 1193670528 M * tanjix hekki everybody 1193670534 M * tanjix s/hekki/hello 1193670539 M * daniel_hozac hi 1193670562 Q * julien Ping timeout: 480 seconds 1193670566 M * tanjix daniel_hozac: i have some trouble hehe, maybe you can give me a hint 1193670571 Q * balbir Ping timeout: 480 seconds 1193670573 M * daniel_hozac what's up? 1193670583 M * tanjix on the new 2.3.0.28 tools it seems that each gust gets a "real" lo-device, right? 1193670598 M * tanjix lo was not there before with older tools 1193670637 M * daniel_hozac s/tools/kernel/ 1193670638 M * daniel_hozac yes. 1193670669 M * tanjix ok, and from this point on, ppl cannot do connects like "telnet localhost 110" 1193670679 M * tanjix telnet EXTERNAL-IP 110 from the vps instead works fine 1193670693 M * tanjix before the lo device came up, everything was fine 1193670694 M * daniel_hozac disable SINGLE_IP. 1193670701 M * tanjix where? 1193670757 M * daniel_hozac either in the kernel as CONFIG_VSERVER_AUTO_SINGLE, or per guest in /etc/vservers//nflags. 1193670819 M * tanjix daniel_hozac: thx, i do not have the nflag file 1193670823 M * tanjix shall i create it ? 1193670830 M * daniel_hozac as usual, yes. 1193670846 M * tanjix and what to put here? SINGLE_IP=no ? 1193670886 M * daniel_hozac ~SINGLE_IP 1193670913 M * tanjix with the tilde ? 1193670930 M * daniel_hozac yes. 1193670934 J * Yv1 yvonne@vpn-9-052.rz.uni-mannheim.de 1193670993 P * Yv1 1193671000 M * tam I'm feeling dumb, and haven't been able to find a wiki page.. Is it possible to mount a cifs share inside a guest? 1193671002 M * tanjix daniel_hozac: thx - IT WORKS!!! you saved me hehe 1193671025 M * daniel_hozac you're welcome! 1193671029 M * daniel_hozac tam: no. 1193671052 M * tam crap. Can I mount it outside the vserver and make it available inside? 1193671060 M * daniel_hozac tam: CIFS creates kernel threads when you mount something, and those are disallowed. 1193671067 M * gebura yes you can 1193671070 M * daniel_hozac tam: sure, just add it to /etc/vservers//fstab.remote 1193671072 Q * esa Ping timeout: 480 seconds 1193671092 M * gebura daniel_hozac, with another implementation (!= samba) , like fuse (userland) can't that be done ? 1193671140 M * daniel_hozac maybe, with the appropriate device nodes and capabilities... 1193671258 M * daniel_hozac but in general, allowing guests to mount things is dangerous. 1193671347 M * gebura i agree 1193671367 J * balbir ~balbir@59.178.33.134 1193672209 Q * Primus_ 1193672252 M * tanjix what was the link to the 2.3.0.28 patch? 1193672259 M * tanjix don*t see it on the wabpage :( 1193672286 M * daniel_hozac http://vserver.13thfloor.at/Experimental/ 1193672302 M * tanjix thx 1193673936 Q * balbir Read error: Operation timed out 1193674232 J * JonB ~NoSuchUse@kg1-98.kollegiegaarden.dk 1193674643 M * toidinamai With vcontext I can run a program in another context, right? 1193674667 M * gebura yes 1193674696 M * toidinamai Does this automatically drop capabilities? 1193674704 M * toidinamai If so how do I prevent that? 1193674725 J * balbir ~balbir@59.178.38.176 1193674816 M * gebura --personality-flags + 1193674816 M * gebura ... set special flags for the given execution domain 1193674831 M * gebura vcontext --help 1193674851 M * gebura i am not sure thats flags = caps 1193675066 M * daniel_hozac toidinamai: contexts do not drop privs by default. 1193675072 M * daniel_hozac toidinamai: you use vattribute for that. 1193675098 M * gebura so what are "flags" ? 1193675149 M * daniel_hozac the option you listed is for the personality, e.g. linux32. 1193675188 M * toidinamai daniel_hozac: Why doesn't "vcontext ... chroot ..." work? 1193675197 M * daniel_hozac hmm? 1193675226 M * toidinamai I get a permission denied error. 1193675244 M * toidinamai vcontext --migrate --xid 6 chroot /var/lib/vservers/cthulhu 1193675244 M * toidinamai chroot: cannot change root directory to /var/lib/vservers/cthulhu: Permission denied 1193675301 M * daniel_hozac you have a barrier set on /var/lib/vservers, no? 1193675322 M * toidinamai I have. 1193675326 M * daniel_hozac cd /var/lib/vservers/cthulhu; vcontext --migrate --xid 6 --chroot -- ... 1193675337 Q * gebura Quit: Quitte 1193675415 M * daniel_hozac note that migrating does drop all privs that the context you're migrating into does not have. 1193675429 Q * lilalinux_ Remote host closed the connection 1193675471 M * toidinamai Hmpf. 1193675491 M * toidinamai I want to run autodir in a vserver. 1193675498 M * daniel_hozac autodir? 1193675506 M * toidinamai autodir uses autofs to create directories on demand. 1193675511 M * toidinamai http://www.intraperson.com/autodir/ 1193675565 M * daniel_hozac i doubt that's going to work... 1193675736 M * toidinamai Maybe with vnamespace? 1193675757 M * daniel_hozac then it runs on the host... 1193675805 M * toidinamai That's ok. 1193675822 M * toidinamai As long as the effect is visible in the guest. 1193675870 M * daniel_hozac vnamespace should be fine then. 1193675882 J * Punkie ~Punkie@melkor.coolhousing.net 1193676147 M * toidinamai Darn it. It doesn't work. 1193676170 M * toidinamai I guess, I'll have to patch pam_mkhome, then. 1193676398 J * esa ~esa@ip-87-238-2-45.adsl.cheapnet.it 1193676403 N * esa eSa| 1193676501 J * fatgoose ~samuel@76-10-149-199.dsl.teksavvy.com 1193676596 M * toidinamai daniel_hozac: Thanks for your help anyways. 1193676918 J * meandtheshell ~markus@85.127.111.62 1193677234 J * fatgoose_ ~samuel@76-10-149-199.dsl.teksavvy.com 1193677295 Q * fatgoose_ 1193677585 Q * fatgoose Ping timeout: 480 seconds 1193677849 J * phedny ~mark@ip56538143.direct-adsl.nl 1193678360 J * fatgoose ~samuel@76-10-149-199.dsl.teksavvy.com 1193678525 Q * fatgoose Remote host closed the connection 1193678556 J * fatgoose ~samuel@76-10-149-199.dsl.teksavvy.com 1193678588 J * bonbons ~bonbons@2001:960:7ab:0:20b:5dff:fec7:6b33 1193678649 Q * derjohn_mobil Ping timeout: 480 seconds 1193678945 Q * phedny Quit: maintenance 1193679448 J * hparker ~hparker@linux.homershut.net 1193679747 J * lilalinux nas@dslb-084-059-002-074.pools.arcor-ip.net 1193679768 J * phedny ~mark@ip56538143.direct-adsl.nl 1193680069 J * janejane ~janeforT@ANantes-257-1-140-170.w90-32.abo.wanadoo.fr 1193680193 Q * janejane 1193680770 Q * pmenier_off Quit: Konversation terminated! 1193680969 Q * ema Quit: leaving 1193681439 N * Bertl_oO Bertl 1193681445 M * Bertl evening folks! 1193681453 M * bXi evening 1193681529 M * bXi Bertl: i had this weird kernel panic the other day 1193681580 M * bXi wasnt vserver related luckily 1193681691 Q * dowdle Remote host closed the connection 1193681865 J * maria29 ~maria29@d033.dhcp212-198-248.noos.fr 1193681975 Q * maria29 1193682029 M * Bertl bXi: what was it about? 1193682155 Q * toidinamai Quit: Leaving 1193682290 Q * mire Ping timeout: 480 seconds 1193682439 M * bXi some lan stuff 1193682641 Q * tanjix 1193682669 J * mire ~mire@15-171-222-85.adsl.verat.net 1193683036 Q * Pazzo Quit: ... 1193684086 Q * Punkie Quit: Odcházím 1193684369 Q * lilalinux Remote host closed the connection 1193684754 J * deltron3k ~zygut@cassowary.riseup.net 1193684772 M * Bertl welcome deltron3k! 1193684788 M * deltron3k thank you Bertl 1193684891 M * deltron3k I'm trying to setup some limits because some vservers are periodically taking up load, impacting others that are behaving nicely, but I am confused about how you do it 1193684948 M * Bertl http://linux-vserver.org/Resource_Limits 1193684955 M * Bertl http://linux-vserver.org/CPU_Scheduler 1193685096 M * deltron3k i have been reading those, and some posts on the mailing lists, but i'm still confused 1193685113 M * Bertl let's hear then ... 1193685215 J * toidinamai ~frank@svenska.toidinamai.de 1193685228 M * Bertl welcome toidinamai! 1193685299 M * deltron3k I've been watching the RSS and VM of each vserver to try and see if one is hogging on of those, but when the load goes up on the host, it doesn't appear to be associated with any spike in either of those resources 1193685341 M * Bertl load is related to processes being willing to run, but not being able to 1193685361 M * deltron3k i noticed that I entered swap heavily last time load went up 1193685370 M * deltron3k I dont know if that is the case for each occurance 1193685398 M * Bertl could be, trashing always raises the load 1193685557 M * deltron3k but I dont see any vserver's RSS or VM go up when the swapping happens 1193685575 M * deltron3k i would expect that swap would happen when memory is exhausted by one of the guests 1193685579 J * julius_ ~julius@p57B24FF5.dip.t-dialin.net 1193685611 M * Bertl deltron3k: no, swapping happens when the host system runs out of memory 1193685691 M * Bertl s/host/whole system/ 1193685739 M * deltron3k sure that is corret 1193685750 M * deltron3k but my host system doesn't do anything, only vserver 1193685779 M * deltron3k so all the guest's memory added together exhaust the memory of the system 1193685794 M * Bertl correct 1193685806 J * dna ~dna@56-246-dsl.kielnet.net 1193685850 M * deltron3k ok, so if that is correct, and I am watching the RSS and VM of each vserver, and none of them is going up in any way 1193685857 M * deltron3k but suddenly I am swapping... 1193685943 M * deltron3k then, what am I missing here? 1193685968 M * Bertl I see a bunch of possibilites, for example 1193685990 M * Bertl - your guests are over the available memory (in total RSS) 1193686004 M * Bertl - most tasks are not active (so it's fine to swap them out) 1193686022 Q * Julius Ping timeout: 480 seconds 1193686027 M * Bertl - something (e.g. crontab) causes them to swap in at once 1193686045 M * Bertl the result, without any change in VM, is trashing 1193686060 M * Bertl (and as a direct result, a raising load) 1193686253 M * deltron3k another interesting data point I find: number of mysql threads skyrockets at this time, also slow query time and querys / second go way up 1193686348 M * Bertl I think you are just observing statistic fluctuations, which cause the overloaded system to trash at some points 1193686365 M * Bertl have you checked that your RAM can actually handle the workset 1193686474 M * deltron3k it could be, each guest is a vserver running different web services, so it also could be a bot crawling causing many database connections open 1193686487 M * deltron3k how can I check that my RAM can handle the workset? 1193686500 M * Bertl check the RSS of all guests, sum it up 1193686520 M * Bertl the perfect setup would be with Sum(RSS) <= RAM 1193686530 M * Bertl but some overbooking is probably fine 1193686749 M * deltron3k if I sum up the current RSS its 379.99M, which is not very much 1193686767 M * deltron3k considering my ram is 2076260 1193686797 M * Bertl that should never cause any swapping then 1193686842 M * Bertl how did you sum the RSS? 1193686856 M * deltron3k i'll try summing the maximum seen from /proc/virtual/#/limit maybe 1193686875 M * deltron3k maximum seen is the second column? 1193686918 M * Bertl yep min/max on newer systems 1193686933 M * deltron3k i was taking the first column number for RSS, mulitplying by 4096 1193686947 M * deltron3k for each number in /proc/virtual/# and then adding those numbers together 1193686948 M * Bertl that's the current set, yes 1193686960 M * deltron3k i'll do the maximum one, i also may have made a calculation error 1193686967 M * Bertl okay 1193687013 J * toidinamai_ ~frank@svenska.toidinamai.de 1193687161 Q * toidinamai Ping timeout: 480 seconds 1193687374 M * deltron3k ok, I took each vurrent RSS value, multipled it by 4096 due to page size, then added those numbers together to get 369209176 1193687402 M * deltron3k that seems really large, is that bytes? 1193687435 M * Bertl yep, that would be bytes 1193687436 M * deltron3k i think if I divide by 1024, then I get the same units I get for total memory presented by 'free', correct? 1193687472 M * Bertl but 400M is still not much for 2GB system 1193687475 M * Bertl +a 1193687496 M * Bertl what kernel version/patches do you use? 1193687528 J * larsivi ~larsivi@101.84-48-201.nextgentel.com 1193687541 M * deltron3k let me do the maximum seen value column 1193687557 M * deltron3k i am using the debian stable 2.6.18-4-vserver-686 1193687615 M * Bertl an update there wouldn't hurt, but should not be related 1193687723 M * deltron3k the maximum column added together is 2210291712 1193687751 M * deltron3k which is over the maximum memory of the system 1193687760 M * deltron3k which does seem to correlate the swap that happens 1193687760 M * Bertl here you got, it seems your guests cause spikes over the physical memory 1193687786 M * Bertl the big questions we cannot answer yet are: 1193687801 M * Bertl - is the maximum reached at the same time? 1193687808 M * deltron3k ok, then my measurements before are not fine grained enough to see these spikes, the measurements are every 5 minutes, so it must be happening in there 1193687829 M * Bertl - what causes/triggers the increase by a factor of 10 1193687908 M * deltron3k and what can I do to keep it from happening! 1193687950 M * Bertl yeah, did your guests hit any limit (last column) 1193687969 M * deltron3k i have no limits set, thats what I am trying to figure out now 1193687976 M * deltron3k how and what I should set them to 1193687993 M * Bertl okay, the maximum reached so far should give you an idea 1193688075 M * Bertl also if you want to take preventive measures, you could reduce the number of maximum threads for services like httpd or mysql 1193688079 M * deltron3k here is what I see now: http://pastebin.com/m48710f0c 1193688213 J * gigi31 ~gigi31@ANantes-257-1-140-170.w90-32.abo.wanadoo.fr 1193688214 M * deltron3k i dont know what is recommended way, but i guess i want to set RSS limits for guests to a number that is some number lower than what I see in the max column? 1193688251 M * Bertl well, that depends on what you are up to :) 1193688279 M * Bertl if you _want_ those guests to work properly, you definitely want to give something above the max observed 1193688321 M * Bertl otherwise the OOM killer will start reaping on that guest at some point 1193688342 Q * gigi31 1193688397 M * deltron3k but if I set a value above the max observed, then what is happening now, will just continue, wont it? 1193688446 M * Bertl correct, moving heavy guests to a different system or expanding on the memory would solve the issues though 1193688463 M * deltron3k ok, so setting resource limits will just cause instability 1193688471 M * deltron3k in the applications running in the guests 1193688499 M * Bertl that depends on the limits 1193688526 M * Bertl some of the limits give immediate response and are handled gracefully in most apps 1193688559 M * Bertl other limits (like memory) cannot provide direct feedback and naturally cannot be handled gracefully 1193688600 M * Bertl if you run a guest using 200MB ram on a 64M system, it will either trash or (if no swap enabled) OOM kill 1193688671 M * Bertl ~ 1193688745 M * deltron3k if I set a memory resource limit on a guest and it reaches that limit, does it swap? or does OOM happen? 1193688768 M * Bertl if you reach the VM limit, the request is simply denied 1193688795 M * Bertl if you rech the RSS limit, then nothing happens immediately, but when memory gets low 1193688811 M * Bertl then the OOM kill is going to kill off tasks from that guest 1193688947 M * mjt . o O { overcommit_memory } 1193688964 M * Bertl hehe, yeah, disabling that would actually help a lot 1193688974 M * deltron3k what does that mean? 1193688981 M * Bertl but I doubt you will be able to start a single guest without overcommitment 1193688997 M * mjt hmm? 1193689077 M * Bertl deltron3k: could you show me the VM maximums for your guests too? 1193689147 M * mjt # cat /proc/sys/vm/overcommit_memory 1193689147 M * mjt 0 1193689147 M * mjt # vserver-stat | wc -l 1193689147 M * mjt 15 1193689150 M * mjt err 1193689156 M * mjt wrong server :) 1193689174 M * mjt # cat /proc/sys/vm/overcommit_memory 1193689174 M * mjt 2 1193689174 M * mjt # vserver-stat | wc -l 1193689174 M * mjt 8 1193689195 M * Bertl that's impressive 1193689198 M * deltron3k Bertl: http://pastebin.com/m333ddab2 1193689224 M * mjt free 1193689225 M * mjt total used free shared buffers cached 1193689225 M * mjt Mem: 510316 503960 6356 0 29632 164068 1193689225 M * mjt -/+ buffers/cache: 310260 200056 1193689225 M * mjt Swap: 995896 12 995884 1193689273 M * mjt but those are 7 rather small servers (running single services like ntp, named, ...) 1193689274 M * deltron3k how does this overcommit_memory you speak of work? 1193689293 M * Bertl basically it's like this in Linux (by default) 1193689301 M * mjt $kernel_source/Documentation/filesystems/proc.txt -- search for overcommit_memory 1193689311 M * Bertl an app asking for 2GB of memory, get's it. period. 1193689331 M * Bertl then, when that app actually starts using up the memory, the RSS rises 1193689346 M * Bertl till there is no memory left, at which point the OOM killer strikes 1193689370 M * Bertl if you disable overcommitment, then the 'requested' memory has to be available for a request to succeed 1193689398 M * Bertl note that e.g. forking will immediately request the same amount of memory already in use 1193689414 M * deltron3k disabled is the default? 1193689427 M * mjt enabled is the default 1193689457 M * Bertl i.e. the overcommitment is the default 1193689473 M * mjt (with the default value "0" being "enabled" :) 1193689498 M * deltron3k i dont mind random things being killed in all the guests, *except* the mysql one... i think that would be dangerous 1193689506 M * deltron3k the other guests are just running apache and cgis 1193689519 M * mjt you can tune it somewhat 1193689530 M * Bertl recent kernels have an OOM bias per guest 1193689544 M * Bertl you could use that as well 1193689564 M * deltron3k i'm not really clear what I should do next 1193689577 M * Bertl restrict the non mysql guests 1193689586 M * Bertl leave the mysql one unrestricted 1193689594 M * deltron3k set the RSS resource limits for those? 1193689607 M * mjt and tune mysql process limits too, i'd guess 1193689653 M * mjt because if your mysql - unrestricted - processes will be killed due to some wild behavour, it'd be... not nice. 1193689654 M * deltron3k i am looking at the mysql parameters separately, its a good parallel thread to investiage 1193689662 M * Bertl yes, limiting threads in the httpd and/or mysql would probably help 1193689701 M * deltron3k so, I should look at setting RSS limits for the apache guests, only RSS? Not VM too? 1193689740 M * Bertl you can do VM too, but that can probably be solved easier with adjusting the apache configs 1193689788 M * deltron3k ok, i will do RSS, look at apache and mysql configs. do you have a recommendation based on my max RSS numbers what I should choose for limits? 1193689800 M * deltron3k should I pick some percentage less than the maximum I have seen? 1193689836 M * Bertl the rule of the thumb is twice the current value 1193689911 M * deltron3k ok, just to make sure, the number I put in /etc/vservers/myguest/rlimits/rss is the same unit as I see in /proc/virtual/myguest/limit RSS row? 1193689933 M * deltron3k so i take what is current now for that guest, and *2 and then put it in that file 1193689944 M * Bertl it should be, although we added support for using 1k instead of page size 1193689962 M * Bertl but you should be able to verify that easily with a single guest 1193689970 M * Bertl (just check with proc) 1193690019 M * daniel_hozac Bertl: we did? 1193690031 M * Bertl daniel_hozac: didn't we? 1193690049 M * Bertl maybe it got lost from my _vary_ long todo list :) 1193690090 M * daniel_hozac if we did add it, i don't see how it's chosen. 1193690114 M * deltron3k it appears to be sames 1193690129 M * daniel_hozac yeah, it's still page-size. 1193690587 J * marcfiu ~mef@aegis.CS.Princeton.EDU 1193691010 Q * meandtheshell Quit: Leaving. 1193691174 M * deltron3k ok, I set some RSS limits, now I will see what happens to this 1193691189 M * deltron3k Bertl: should I be setting a scheduluer? would that help me? 1193691219 M * daniel_hozac only if you want to limit CPU usage. 1193691228 M * daniel_hozac or, rather, control it. 1193691388 P * marcfiu 1193691462 M * deltron3k i think my problem has to do with sawp, which creates high cpu 1193691466 M * deltron3k so maybe I do not want that 1193691617 Q * baggins Ping timeout: 480 seconds 1193691839 J * Aiken ~james@ppp121-45-228-65.lns1.bne4.internode.on.net 1193691856 J * baggins ~baggins@kenny.mimuw.edu.pl 1193691911 Q * JonB Quit: This computer has gone to sleep 1193692217 M * Bertl wb Aiken! baggins! 1193692351 M * Aiken hello 1193692546 M * deltron3k ok, thanks for your help Bertl!! 1193692549 M * deltron3k bye now 1193692550 Q * deltron3k Quit: leaving 1193692769 Q * bonbons Quit: Leaving 1193693406 Q * julius_ Remote host closed the connection 1193693592 M * daniel_hozac Bertl: http://svn.linux-vserver.org/projects/util-vserver/browser/trunk/scripts/vmount :) 1193693723 M * Bertl cool! 1193693825 M * Bertl daniel_hozac: I was wondering ... regarding the CoW revalidation, cound't we do that in a similar way the rename handles it? 1193694247 M * daniel_hozac Bertl: how does rename handle it? 1193694302 M * Bertl do_rename() does the old_dentry = lookup_hash(&oldnd); 1193694485 M * daniel_hozac thus ensuring it's still valid? 1193694527 M * Bertl well, I thought more that we could use this in the 'redo' case 1193694549 M * Bertl although I haven't tried it yet .. IMHO the new entry should be in the hash by then, no? 1193694576 M * daniel_hozac ah, yes. 1193694628 M * daniel_hozac and if it's not, lookup_hash performs the lookup. 1193694690 M * Bertl I also think that this should not need to free up the nameidata 1193694740 M * daniel_hozac right. 1193694783 M * daniel_hozac we'd have to use LOOKUP_PARENT though. 1193694808 J * Yvo ~yvonne@91.64.217.106 1193694820 M * Bertl wb Yvo! 1193694830 M * Bertl daniel_hozac: was just an idea ... 1193694843 M * Yvo thx! 1193694878 M * fb_ time to sleep, good night 1193694887 M * Bertl have a good one! 1193694908 Q * larsivi Quit: Konversation terminated! 1193695357 Q * DuckMaster Remote host closed the connection 1193695572 J * duckx ~Duck@81.57.39.234 1193695603 Q * balbir Read error: Operation timed out 1193695955 J * marcdirix ~marc@86.85.244.215 1193696001 Q * dna Quit: Verlassend 1193696302 Q * baggins Ping timeout: 480 seconds 1193696332 J * balbir ~balbir@59.178.39.209 1193696353 P * marcdirix 1193696810 J * baggins ~baggins@kenny.mimuw.edu.pl 1193696958 J * almak ~almak@willers.employees.org 1193696973 M * almak hi 1193696976 M * Bertl wb almak! 1193696982 M * almak hey bertl 1193697033 M * almak question for you. If I want to perform resource limit on processes running in the host. Is ther a way I can assign a vserver for it? 1193697052 M * Bertl yes and no :) 1193697063 M * almak ok? 1193697069 M * Bertl you cannot limit host processes with Guest limits 1193697086 M * Bertl but you can put arbitrary services into a context, and then limit them 1193697122 M * almak will vserver util allow me to do this? 1193697127 M * Bertl daniel_hozac: do you have a patch at hand for the CoW fixes you tested? 1193697143 M * Bertl almak: yes, util-vserver can do that for you 1193697144 M * daniel_hozac Bertl: hmm? 1193697161 M * daniel_hozac Bertl: IIRC you wrote all of the patches :) 1193697161 M * Bertl the put of the nameidata 1193697169 M * daniel_hozac ah. 1193697194 M * almak how do I mix apps running in guest and host processes in a context? 1193697228 M * Bertl what do you want to mix there? 1193697311 M * almak I have apps running in a proper guestOS but I have some services on the host side that I do want the CPU to spike. How do I setup vserver utils for the host services and the guest running at the same time? 1193697330 M * almak I do not want CPU to spike (I mean) 1193697344 M * Bertl why not simply put the services into a new guest? 1193697444 M * almak It needs to access a number of files scattered around on the host side and also has some shared memory stuff access going on with the host. Want to minimize code change on the host side. 1193697472 M * daniel_hozac Bertl: http://people.linux-vserver.org/~dhozac/p/k/delta-cow-fix21.diff 1193697482 M * Bertl almak: i.c., well, just putting it into a context should be sufficient 1193697507 M * Bertl almak: note that shared memory can be tricky as you have to share IPC namespaces with the host in this case 1193697540 M * daniel_hozac (which is not currently supported by the utils) 1193697542 M * Bertl daniel_hozac: tx, the unlock_new keeps surfacing, or is that a real fix this time? 1193697562 M * daniel_hozac it's unrelated, but it's a fix nonetheless. 1193697580 M * almak yes. So to put the process in a context do I use vcmd to start the process? 1193697590 M * daniel_hozac yes, that would work. 1193697612 M * almak how do I set the CPU limit using vcmd? 1193697627 M * Bertl you can probably use the tools for that 1193697663 M * almak ok. I will take a look. Thx 1193697677 M * Bertl you're welcome! 1193697745 J * derjohn_mobil ~aj@e180204105.adsl.alicedsl.de 1193697793 M * Bertl wb derjohn_mobil! :) 1193697889 M * derjohn_mobil Hello Bertl! 1193698449 J * click_ click@ti511110a080-4188.bb.online.no 1193698562 Q * click Ping timeout: 480 seconds 1193698649 P * click_ 1193698733 J * yarihm ~yarihm@84-75-130-73.dclient.hispeed.ch 1193699024 Q * derjohn_mobil Ping timeout: 480 seconds 1193699117 J * derjohn_mobil ~aj@e180223153.adsl.alicedsl.de 1193699592 Q * yarihm Quit: Leaving 1193701959 Q * balbir Read error: Operation timed out 1193702136 J * friendly12345 ~friendly@ppp59-167-134-217.lns3.mel6.internode.on.net 1193702153 M * Bertl wb friendly12345! 1193702380 M * Yvo good night! 1193702390 M * Bertl night Yvo!