1234051297 M * Supaplex I don't see where the wiki sheds any light on this. I'll look again after $lunch :) 1234051424 M * Bertl http://linux-vserver.org/Frequently_Asked_Questions#How_do_I_manage_a_multi-guest_setup_with_vhashify.3F 1234051452 M * Supaplex the mount with tag option is where I'm lost 1234051488 M * Bertl well, if you don't care about disk space accounting, you don't need the tagging 1234051491 M * Supaplex How do I tag a guest's directory with xid? 1234051495 M * Supaplex ahh. 1234051504 M * Supaplex yeah, not an issue 1234051509 M * Bertl not the directory, you mount the filesystem with 'tag' as option 1234051532 M * Supaplex this faq entry mentions the -o tag. 1234051539 M * Supaplex but only needed for accounting, right? 1234051568 M * Bertl accounting and security for shared spaces, not required for unification per se 1234051777 Q * bonbons Quit: Leaving 1234052606 Q * mattzerah Read error: Operation timed out 1234052731 Q * geb Ping timeout: 480 seconds 1234052761 Q * hparker Quit: Read error: 104 (Peer reset by connection) 1234053189 J * friendly ~friendly@ppp121-44-207-248.lns10.mel4.internode.on.net 1234053207 J * mattzerah ~matt@121.50.221.124 1234053294 J * geb ~geb@92.140.134.74 1234053381 Q * harobed Ping timeout: 480 seconds 1234053766 Q * mattzerah Read error: Operation timed out 1234054299 J * harobed ~harobed@arl57-1-82-231-110-14.fbx.proxad.net 1234054381 J * mattzerah ~matt@121.50.221.124 1234054847 J * jfs ~jfs@ip-80-236-226-213.dsl.scarlet.be 1234056613 Q * harobed Ping timeout: 480 seconds 1234057231 Q * jfs Quit: Quitte 1234058704 Q * bibabu Remote host closed the connection 1234060843 Q * geb Remote host closed the connection 1234061407 J * bibabu bibabu@vserv1.de 1234062049 J * derjohn_foo ~aj@e180196198.adsl.alicedsl.de 1234062477 Q * derjohn_mob Ping timeout: 480 seconds 1234065196 Q * duckx Remote host closed the connection 1234065637 Q * mattzerah Read error: Operation timed out 1234065794 M * Bertl off to bed now ... have a good night! 1234065799 N * Bertl Bertl_zZ 1234066433 J * mattzerah ~matt@121.50.221.124 1234067335 Q * friendly Quit: Leaving. 1234071556 Q * mattzerah Remote host closed the connection 1234080413 J * kiorky_ ~kiorky@cryptelium.net 1234080446 Q * kiorky Remote host closed the connection 1234082265 J * doener_ ~doener@i577BBA15.versanet.de 1234082371 Q * doener Ping timeout: 480 seconds 1234083320 J * friendly ~friendly@ppp121-44-207-248.lns10.mel4.internode.on.net 1234083664 Q * nenolod Ping timeout: 480 seconds 1234083762 J * nenolod nenolod@petrie.dereferenced.org 1234084864 Q * nenolod Ping timeout: 480 seconds 1234085242 J * nenolod nenolod@petrie.dereferenced.org 1234086830 Q * friendly Quit: Leaving. 1234087066 Q * larsivi Remote host closed the connection 1234087081 J * larsivi ~larsivi@9.80-202-30.nextgentel.com 1234087590 J * yarihm ~yarihm@guest-docking-nat-1-005.ethz.ch 1234089823 J * cga ~weechat@94.36.93.49 1234090261 Q * dodo Remote host closed the connection 1234090282 J * kafyfodu ~kafyfodu@193.43.249.169 1234091226 Q * kafyfodu Remote host closed the connection 1234091290 J * cucuko ~cucuko@193.43.249.169 1234091794 M * derjohn_foo Bertl_zZ, did you write CPU_Scheduler ? Thanks for the clarification! My kernel is running now (after several troubles with debian and xen domU support in the kernel). I'll now try to set some limits. (without limits set the kernel runs fine) 1234092393 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d 1234093398 J * geb ~geb@92.140.213.254 1234093423 M * geb hi 1234094277 J * saulus ~saulus@c189095.adsl.hansenet.de 1234094387 Q * saulus_ Ping timeout: 480 seconds 1234094735 J * duckx ~Duck@81.57.39.234 1234098374 Q * cga Quit: WeeChat 0.2.6 1234098881 Q * Aiken Remote host closed the connection 1234099400 J * cga ~weechat@94.36.93.49 1234102563 Q * nou Remote host closed the connection 1234103922 N * ensc Guest409 1234103932 J * ensc ~irc-ensc@p57AA7C87.dip.t-dialin.net 1234104043 Q * Guest409 Ping timeout: 480 seconds 1234104224 J * nou Chaton@causse.larzac.fr.eu.org 1234104783 J * saulus_ ~saulus@c195187.adsl.hansenet.de 1234104786 J * mrjack vend3r@office.smart-weblications.net 1234104789 M * mrjack hi all! 1234104819 M * mrjack how can i make reboot work within a vserver - i setup init-style as plain, when i reboot or reboot -f the vserver just stops, but does not start again.. any hints? 1234104892 Q * saulus Ping timeout: 480 seconds 1234106890 N * Bertl_zZ Bertl 1234106894 M * Bertl morning folks! 1234106922 M * Bertl mrjack: a reboot from inside with plain initstyle is not really easy 1234106951 M * Bertl you need to enable the reboot helper and act properly 1234106978 M * Bertl and init has to call sys_reboot() before terminating 1234107069 M * Bertl off for now, bbl ... 1234107074 N * Bertl Bertl_oO 1234107169 Q * mrjack Read error: Connection reset by peer 1234107494 Q * cucuko Remote host closed the connection 1234107515 J * sufe ~sufe@193.43.249.169 1234107568 J * mrjack_ vend3r@office.smart-weblications.net 1234110348 Q * ard Ping timeout: 480 seconds 1234110762 J * xdr_ ~xdr@gote2.247.cust.blixtvik.net 1234111519 M * saulus_ hi Bertl_oO - as I noticed, I have to patch the vservers into the vanilla. I got linux-2.6.28.4 and your patch, but what version of vservers shall I take? Is the stable one vs2.2.0.7 compatible with my kernel? 1234111595 M * pmjdebruijn no 1234111599 M * pmjdebruijn there are experimental patches... 1234111615 M * pmjdebruijn saulus_: I recommend you stick with 2.6.27.x because that will be long-term maintained... 1234111635 M * pmjdebruijn but even 2.6.27 with vserver is still experimental 1234111650 M * pmjdebruijn saulus_: if you need absolute rock-solid-ness stick with 2.6.22.19-vs2.2.0.7 1234111681 M * saulus_ pmjdebruijn: Bertl_oO wrote a patch for me: http://vserver.13thfloor.at/Experimental/delta-fssec-feat01.diff - I dont know, if this works with older kernels 1234111695 M * pmjdebruijn sorry, no clue 1234111771 M * saulus_ I will wait here until he answears (probably in a short time) before making my server crash :) 1234111903 M * saulus_ probably it wont work wirh 2.6.27 because of this line in the patch: diff -NurpP --minimal linux-2.6.28.3-vs2.3.0.36.5 1234112321 M * derjohn_foo saulus_, well, it is probably for devel (2.3....) 1234114680 M * derjohn_foo Bertl_oO, I am compiling the kernel within an "buckt-limited" guest and have a load AVG of about 0.25. To: It looks like working. It didn try two concurrent guests, though. 1234114885 M * derjohn_foo Bertl_oO, oh oh ... I just recognize I am compiling in a non-limited guest. Did you set schedule defaults ? 1234115672 Q * geb Quit: Quitte 1234115690 J * Pazzo ~ugelt@reserved-225136.rol.raiffeisen.net 1234116348 M * saulus_ do you think its problematic, when I patch the 2.6.28.4 kernel with the 2.6.28.3 patch? 1234116653 M * cehteh probably not, you'll see where it conflicts 1234116839 M * saulus_ what should the patch command be like? Im trying from /usr/src, files having there, kernel unpacked, link to "./linux": cat patch-2.6.28.3-vs2.3.0.36.5.diff | patch -p0 2>&1 | tee patch.ou 1234116895 A * cehteh usually uses patch -p1 1 hr. I just thought I could save time ... 1234133700 M * derjohn_foo OTOH a cache sounds dangerous, if I change something afterwards (config option, patch ....) 1234133718 M * Bertl nah, ccache handles that properly 1234133746 M * arekm the first idea should be building on a faster machine (if available) 1234133753 M * derjohn_foo so I can work "as usual", just must have ccache installed and everything is faster after the 1st try? 1234133760 M * Bertl but, if you want to cut down on the kernel build time, the big trick is just to compile stuff you _need_ not everything 1234133806 M * Bertl ccache, distcc, all might speed up your build 1234133807 M * derjohn_foo well, as I tend to make distro-like kernels as debian packages, stripping down would be the wrong approach. 1234133832 M * Bertl well, IMHO distro like kernels _are_ the wrong approach per se :) 1234133844 M * derjohn_foo *lol* 1234133858 M * arekm hehe 1234133866 M * Bertl no real advantage, and a number of disadvantages 1234133876 M * derjohn_foo let's say: i want on packerl for all machines we maintain. 1234133890 M * Bertl basically the only reason to build distro like kernels is if you maintain a distribution :) 1234133897 M * arekm there is one huge advantage - exactly the same binary run on all machines. Easier debugging/maintaining/verifying and so on 1234133936 M * derjohn_foo but I agree that i compile many many modules I never used. 1234133947 M * derjohn_foo (and probably will never use) 1234133957 M * Bertl and a bunch of features which actually harm your kernel :) 1234133999 M * derjohn_foo like my 300Hz approach :) ? 1234134023 M * Bertl well, I would opt for highmem :) 1234134052 M * arekm some of us are in 64bit world already ;-P 1234134055 M * derjohn_foo in what way? x86_64 doesnt have any problems with that? 1234134072 M * Bertl so all 64bit already? 1234134084 M * derjohn_foo on x86, I made the observation that PAE doesnt run on all HW 1234134107 M * derjohn_foo well, yes, the hosts are 64bit, but we run many guests 32bit 1234134131 M * arekm derjohn_foo: only host kernel matters in this case 1234134155 M * derjohn_foo and until now i didnt really understand what the memory split is for. 1234134192 M * Bertl mainly for avoiding highmem and the related overhead 1234134240 M * derjohn_foo what why 2/2 or 3/1 ? whats the difference ? 1234134269 M * Bertl how much address space is reserved for the kernel 1234134319 Q * arekm Quit: reboot (2.6.28.4 + grsecurity full + apparmor + vserver test) 1234134321 M * derjohn_foo does the kernel need more than 1GB to address? 1234134336 M * Bertl if you want to use more than 1GB then yes 1234134369 M * derjohn_foo and the 3/1 split? means the kernel gets one 1 Gb and processes 3gb ? 1234134375 N * ensc Guest484 1234134385 J * ensc ~irc-ensc@p57AA7C87.dip.t-dialin.net 1234134398 J * FCOJ ~mordur@dsl-195-100.hive.is 1234134492 Q * Guest484 Ping timeout: 480 seconds 1234134573 A * derjohn_foo reads up http://kerneltrap.org/node/2450 1234134583 M * Bertl 3/1 is 3GB user, 1GB kernel 1234134602 M * Bertl 1/3 is the split you want for up to 3GB of memor 1234134605 M * Bertl +y 1234134617 M * Bertl if you want more memory, you better go for x86_64 1234134701 J * arekm arekm@carme.pld-linux.org 1234135372 M * derjohn_foo Bertl, thanks so far - .4 is still compiling ... 1234135407 M * derjohn_foo i'll tell you my further oberservations, but currently everything looks fine. 1234135424 M * Bertl good! thanks! 1234137354 Q * duckx Remote host closed the connection