1220316532 Q * esa` Ping timeout: 480 seconds 1220317554 J * esa bip@62.123.8.106 1220318305 J * esa` bip@62.123.8.137 1220318323 Q * esa Ping timeout: 480 seconds 1220323242 J * derjohn_foo ~aj@p5B23DFAA.dip.t-dialin.net 1220323672 Q * derjohn_mob Ping timeout: 480 seconds 1220325523 Q * Genghis Ping timeout: 480 seconds 1220327723 Q * balbir Ping timeout: 480 seconds 1220328387 J * balbir ~balbir@122.167.111.224 1220328883 T * * http://linux-vserver.org/ |stable 2.2.0.7, devel 2.3.0.34, grsec 2.2.0.7|util-vserver-0.30.215|libvserver-1.0.2|vserver-utils-1.0.3| He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the Wiki, and we forget about the minute. 1220328883 T * ChanServ - 1220330854 N * Bertl_zZ Bertl 1220330857 M * Bertl morning folks! 1220333275 J * Genghis ~Genghis@what.is.this.digitalcrap.org 1220333314 N * Genghis Guest4847 1220334061 N * Guest4847 Genghis 1220334657 M * fb morning Bertl :-) 1220335507 Q * quote Ping timeout: 480 seconds 1220335745 Q * larsivi Quit: Konversation terminated! 1220336144 J * sharkjaw ~gab@149-67-194.231210.adsl.tele2.no 1220336789 Q * balbir Ping timeout: 480 seconds 1220338475 J * larsivi ~larsivi@85.221.53.194 1220338621 N * DoberMann[PullA] DoberMann 1220339656 M * torsti76 mornin' bertl 1220339734 J * balbir ~balbir@59.145.136.1 1220339781 M * torsti76 i'm out... 1220339783 P * torsti76 1220340708 J * dna ~dna@p54BCF495.dip.t-dialin.net 1220341644 Q * FireEgl Ping timeout: 480 seconds 1220342279 J * FireEgl FireEgl@173.16.9.10 1220342358 J * quote ~quote@pomoc.ircnet.com 1220342359 M * quote hi 1220342383 M * quote is it possible to stop a vserver somehow without vserver stop command? 1220342391 M * quote like killing a pid?:) 1220342419 M * daniel_hozac vkill. 1220342425 M * Bertl yes, you can do that with vkill 1220342454 M * quote umm... 1220342462 M * quote but it requires a pid or something 1220342476 M * quote No pid specified 1220342481 M * daniel_hozac yes. 1220342490 M * quote how can i see which is the guests pid?:) 1220342507 M * daniel_hozac and, as per the regular kill(2) rules, -1 will kill all processes. 1220342533 M * quote i just want to stop one guest vserver 1220342540 M * quote which i cant, cause it has config in the wrong place 1220342546 M * quote i recompiled util-vserver 1220342559 M * quote but in between i created a guest vserver 1220342566 M * quote and started it 1220342568 M * quote now i cant stop it :p 1220342600 M * quote vserver test stop 1220342601 M * quote Can not find a vserver-setup at '/etc/vservers/test/'. 1220342619 M * quote cause its in /usr/vservers/test 1220342622 M * daniel_hozac vserver /path/to/config stop 1220342626 M * quote oh 1220342653 M * quote vserver 'test' is not running 1220342660 M * quote but i still see it in vserver-stat, lol 1220342665 M * quote 1008 2 10.2M 892K 0m00s56 0m00s40 16h54m55 1220342667 M * quote only with no name 1220342671 M * quote just context and this stuff.. 1220342748 M * Bertl so you want to use vkill in this case, on xid=1008 1220342776 M * Bertl first with -1, then 0 and then 1 as pid (you can stop when the context is gone :) 1220342804 M * quote vkill -1 --xid=1008 1220342805 M * quote ? 1220342813 M * Bertl vkill --help 1220342879 M * quote got it :P 1220342879 M * quote thx 1220342923 M * Bertl np 1220343106 Q * FireEgl Read error: Operation timed out 1220343830 J * FireEgl FireEgl@173.16.9.10 1220343952 Q * larsivi Quit: Konversation terminated! 1220344186 J * larsivi ~larsivi@85.221.53.194 1220344186 Q * FireEgl Read error: Connection reset by peer 1220344527 Q * larsivi Quit: Konversation terminated! 1220344730 J * larsivi ~larsivi@85.221.53.194 1220344873 Q * derjohn_foo Ping timeout: 480 seconds 1220346050 Q * dna Quit: Verlassend 1220347982 Q * larsivi Quit: Konversation terminated! 1220348183 J * larsivi ~larsivi@85.221.53.194 1220348470 J * loddafni1 ~mike@chello080108111137.5.11.univie.teleweb.at 1220349014 J * sono ~sono@78.46.42.175 1220349118 P * sono 1220349263 J * FireEgl FireEgl@173.16.9.10 1220350204 Q * pmenier_off Read error: Connection reset by peer 1220350233 J * pmenier_off ~pme@LNeuilly-152-22-72-5.w193-251.abo.wanadoo.fr 1220350735 Q * larsivi Quit: Konversation terminated! 1220350800 J * larsivi ~larsivi@85.221.53.194 1220351113 Q * esa` Ping timeout: 480 seconds 1220351143 J * esa bip@62.123.8.50 1220352272 Q * larsivi Quit: Konversation terminated! 1220352323 J * larsivi ~larsivi@85.221.53.194 1220352372 Q * larsivi 1220352463 J * larsivi ~larsivi@85.221.53.194 1220352590 Q * FireEgl Read error: Connection reset by peer 1220352671 Q * balbir Ping timeout: 480 seconds 1220352850 Q * larsivi Remote host closed the connection 1220352903 J * larsivi ~larsivi@85.221.53.194 1220352912 Q * larsivi 1220352984 J * larsivi ~larsivi@85.221.53.194 1220353119 Q * Loki|muh Remote host closed the connection 1220353253 J * FireEgl FireEgl@173.16.9.10 1220353556 M * Bertl short nap ... bbl 1220353561 N * Bertl Bertl_zZ 1220354874 Q * larsivi Quit: Konversation terminated! 1220354921 J * larsivi ~larsivi@85.221.53.194 1220354925 Q * larsivi 1220354983 J * larsivi ~larsivi@85.221.53.194 1220355018 Q * larsivi Remote host closed the connection 1220355086 J * _Keks_ ~keks@80.64.184.134 1220355095 M * _Keks_ hi 1220355651 J * larsivi ~larsivi@85.221.53.194 1220356231 J * RenOS ~renos@athedsl-94096.home.otenet.gr 1220356243 M * RenOS hello everyone 1220356248 J * balbir ~balbir@59.145.136.1 1220356653 M * _Keks_ hi 1220357536 Q * Aiken Quit: Leaving 1220357793 J * balbir_ ~balbir@59.145.136.1 1220358139 Q * balbir Ping timeout: 480 seconds 1220358248 Q * pisco Ping timeout: 480 seconds 1220359066 N * Bertl_zZ Bertl 1220359104 M * Bertl back now .. 1220359125 M * Bertl hey _Keks_, RenOS! 1220359133 M * _Keks_ hi bertl 1220359153 M * _Keks_ do you have something new about the vserver-utils? 1220359184 M * Bertl hmm, what do you need there? 1220359192 M * _Keks_ aehm 1220359198 M * _Keks_ the procfs stuff 1220359200 M * _Keks_ :) 1220359265 M * _Keks_ vserver SERVER start fails with: cant access /proc/uptime blaaa 1220359283 M * Bertl did you run vprocunhide (on the host)? 1220359321 M * _Keks_ manually 1220359336 M * _Keks_ no changes 1220359350 M * _Keks_ downgrade of vserver-utils fix it 1220359367 M * Bertl what util-vserver version did you use? 1220359378 J * pisco ~pisco@tor.noreply.org 1220359380 M * _Keks_ debian/lenny 1220359393 M * _Keks_ should by 315-72xx 1220359401 M * _Keks_ should be 315-72xx 1220359414 M * _Keks_ 2sec. 1220359417 M * Bertl 0.30.315 should be fine, even in debian :) 1220359430 M * _Keks_ thats the problem 1220359432 M * Bertl 0.30.216 (pre something) had a bug 1220359436 M * waldi _Keks_: please show the exact version. testing may not be uptodate yet 1220359461 M * _Keks_ hm 1220359464 M * waldi you need 0.30.216~r2772-2 1220359471 M * _Keks_ ic 1220359490 M * _Keks_ ive set up a fresh lenny in a vm 1220359499 M * _Keks_ only vserver-kernel + utils 1220359507 M * _Keks_ vserver build fails 1220359529 M * Bertl again, what versions (kernel, patch, util-vserver :) 1220359543 M * _Keks_ just a sec. 1220359557 M * Bertl take your time ... 1220359575 M * _Keks_ 0.30.216~r2750-3 for the utils 1220359581 M * daniel_hozac get newer ones. 1220359602 M * Bertl yep, 216~r2772-2 is out and should be fine IIRC 1220359841 J * derjohn_foo ~aj@51.42.69.80.in-addr.net-lab.net 1220359858 M * _Keks_ is there a chance to find out which kernel-patch they used? 1220359861 M * _Keks_ hi john 1220360147 M * _Keks_ thanks for your help 1220360179 M * Bertl waldi probably knows how to figure that out 1220360192 M * _Keks_ but he is gone ... 1220360219 M * Bertl but I'm sure you have a kernel with a quite recent API 1220360236 M * Bertl (otherwise you wouldn't have the issues with 0.30.216pre tools) 1220360635 M * _Keks_ :) 1220360652 M * _Keks_ good to know 1220361215 Q * sharkjaw Remote host closed the connection 1220361267 M * waldi _Keks_: the changelog 1220361340 Q * larsivi Quit: Konversation terminated! 1220361414 M * _Keks_ 2.3.0.35 1220361416 M * _Keks_ thnx 1220362788 Q * balbir_ Read error: Operation timed out 1220362949 J * dowdle ~dowdle@scott.coe.montana.edu 1220363440 M * RenOS I'm building a new guest using the rsync method. I need the guest to be synced on an NFS server. When I define the --rootdir /jails/vserver and provide as source a guest installed on the main machine (not the NFS fs) I get the following error: 1220363450 M * RenOS rsync: link_stat "/jails/vservers/vserver2/." failed: No such file or directory (2) 1220363468 M * RenOS if I define the source to be a guest from the NFS mount everything works ok 1220363473 M * RenOS could this be a bug? 1220363537 M * daniel_hozac run the build with --debug, see what's going on... 1220363553 M * RenOS ok daniel 1220363605 Q * _Keks_ Read error: No route to host 1220363638 M * quote wtf 1220363638 M * quote shell-init: error retrieving current directory: getcwd: cannot access parent directories: No such file or directory 1220363644 M * quote when i do vserver enter 1220363680 J * balbir_ ~balbir@59.145.136.1 1220363687 M * Bertl could be some kind of improper barrier or permission 1220363702 M * Bertl but make sure that your directory doesn'T go away when you move 1220363727 M * Bertl i.e. do something like 'pwd' 'cd /' 'cd ' (from the pwd) 1220363803 M * quote ah 1220363804 M * quote cool :p 1220364675 M * RenOS daniel_hozac, rsync does have a point for complaining: http://paste.linux-vserver.org/12413 1220364714 M * RenOS according to this page: http://linux-vserver.org/Building_Guest_Systems I only have to define the virtual server as the source (and not a full path to vdir) 1220364767 N * pmenier_off pmenier 1220364913 M * daniel_hozac whoops, looks like that's only in clone. 1220365067 M * RenOS so, we can conclude that there an error in the document? 1220365105 M * daniel_hozac for now, sure. 1220365527 J * nas ~chatzilla@bb116-14-47-227.singnet.com.sg 1220365536 M * nas hello 1220365592 M * nas hi everyone 1220365619 M * nas anyone tried implementing linux-HA in a vserver guest 1220365630 M * nas please refer to this post for more info 1220365631 M * nas http://forums.debian.net/viewtopic.php?t=30468&highlight= 1220365640 M * daniel_hozac doesn't make much sense to do it inside. 1220365643 M * nas thanks in advance 1220365728 J * laguneeza ~kingarthu@119.95.177.170 1220365733 M * nas ahahah 1220365735 M * nas king 1220365750 M * nas hi daniel 1220365754 M * Bertl welcome nas! 1220365762 M * nas thanks bertl 1220365772 M * laguneeza hi nas 1220365806 P * laguneeza 1220365810 M * nas daniel you think in my setup on the link provided, if i put load balancer with heartbeat in host server, you think that's fine 1220365851 M * RenOS you can only use a load balancer that will also keep track of the available servers 1220365866 M * RenOS so, you can balance the load among the servers and also provide a HA architecture 1220365919 M * RenOS I think nginx can do that 1220365933 M * nas i am using debian though 1220365936 M * ard if you are using debian, installing keepalived will suffice 1220365938 M * nas need to use debain 1220365947 M * nas it's a requirement 1220365950 M * ard hearbeat will make you suffer 1220365959 M * nas have you guys looked into the post? 1220365962 M * ard it's a *good* requirement :-) 1220365970 M * nas http://forums.debian.net/viewtopic.php?t=30468&highlight= 1220365989 M * ard wel yes, but you mistake the concept of vserver with a virtual machine 1220366002 M * nas huh?! 1220366007 M * nas what you mean? 1220366019 M * ard load balancer is a kernel routing networking thing 1220366039 M * ard vserver is a userspace separation thing 1220366044 M * RenOS ard depends on the OSI layer you with to balance 1220366051 M * Bertl ard: depends, e.g. 'balance' is pure userspace 1220366055 M * RenOS nas is talking for a HTTP one 1220366071 M * ard ldirector uses lvs right? 1220366093 M * ard but a userspace balancer might suffice ;-) 1220366139 M * nas Physical Host 1 1220366141 M * nas vserver1 - load balancer1 (ldirectord-2 and heartbeat-2) - load balancer for http requests 1220366142 M * nas vserver2 - web server1 (apache2) - real http server 1220366144 M * nas Physical Host 2 1220366145 M * nas vserver1 - load balancer2 (ldirectord-2 and heartbeat-2) - load balancer for http requests 1220366147 M * nas vserver2 - web server2 (apache2) - real http server 1220366156 M * nas i only have two physical servers 1220366157 M * ard if nas has the resources, he should use keepalived on 2 separate servers, and make loadbalancing/firewalling appliances with it 1220366181 M * nas i need two load balancers and two web servers for high availability and failover 1220366188 M * nas that's why i used vservers 1220366202 M * ard nas : but if ldirectord uses lvs, you can better run it in the host, since it will be changing kernel tables. 1220366216 M * nas coz it would be a waste to dedicate the whole server as just web server for instance 1220366237 M * ard but I would run keepalived in the host, and use the rest in vservers ;-) 1220366295 M * ard heartbeat can give you a major pain to set up. 1220366309 M * nas ldirector uses ipvsadm 1220366326 M * nas not sure about what lvs you mean 1220366333 M * ard keepalived directly talks to the lvs kernel 1220366338 M * ard lvs == ipvs 1220366343 M * ard linuxvirtualserver ;-) 1220366372 M * ard keepalived does a vrrp like failover, and monitors your webservers also... 1220366374 M * nas yeah i know 1220366384 M * nas lvs = vserver right? 1220366390 M * ard no... :-) 1220366407 M * ard lvs == ipvs , http://www.linuxvirtualserver.org 1220366433 M * ard thats the loadbalancing part in the kernel that is controled by ldirector or keepalived 1220366473 M * nas ok 1220366485 M * nas so i have two physical servers, i need two web servers 1220366487 M * ard so you have linuxvirtualserver and linux vserver ;-) 1220366490 M * nas what do you suggest 1220366510 M * nas i also need load balancer to balance requests on two web servers 1220366514 M * ard I would suggest keepalived (since the ultramonkey is too complex for a web thing) on the host 1220366525 M * ard and put your services in vservers 1220366565 M * nas ok ard, clarify with you 1220366589 M * nas you mean i put my load balancer in the host? 1220366593 M * ard If you get a pain from the loadbalancing part, you can use keepalived to only do the failover part, and run a userspaces loadbalancer (like pound) on both systems 1220366616 M * ard nox: the loadbalancer is in the kernel... That's one level deeper :-) 1220366623 M * nas is keepalive in debian etch? 1220366631 M * ard you let keepalived control the loadbalancer 1220366635 M * ard yes 1220366654 M * ard keepalived - Failover and monitoring daemon for LVS clusters 1220366686 M * nas ard i'm lost now 1220366705 M * ard :-) 1220366728 M * nas i still put the web server inside the vserver guest is it? 1220366741 M * ard Yes, but for a reason of stability and security 1220366752 M * ard (environment stability of the host) 1220366787 A * ard must make a coffebreak right now... 1220366791 M * ard sorry... 1220366799 M * nas it's ok 1220366803 M * nas please be back 1220366811 M * nas =) 1220367221 M * nas renos you still there? 1220367223 Q * balbir_ Read error: Operation timed out 1220367245 M * Bertl ard has been around for quite some while ... so don't worry .. patients is the motto :) 1220367289 M * Bertl *patience I mean :) 1220367357 M * nas ahahaha 1220367405 M * nas bertl are you poetzl? 1220367522 M * sid3windr /wii Bertl 1220367547 J * larsivi ~larsivi@169.80-202-217.nextgentel.com 1220367567 M * sid3windr (or /whois) 1220367606 M * Bertl nas: yep 1220367607 M * nas ard by the way just to clarify the load balancer is so that i can use the two web servers simultaneously 1220367615 M * nas yey 1220367619 M * nas hi there 1220367638 M * nas bert thanks for your replies 1220367693 M * nas so that the load balnacer can balance the http requests on the two web servers 1220367737 M * nas the two web servers are in a vserver guest on separate physical servers 1220367813 M * nas that is my high availability part 1220367840 M * Bertl okay 1220367860 M * Bertl if you want to have some kind of failover scenario 1220367886 M * Bertl i.e. so that one guest (on phys host A) can take over when the other one (host or guest) fails 1220367899 M * Bertl you need to put the balancer _outside_ both hosts 1220367940 M * daniel_hozac well, using heartbeat to migrate the IP should work too. 1220367956 M * daniel_hozac but external load balancers make everything smoother... 1220367956 M * Bertl not with a blancer 1220367964 M * Bertl *balancer that is 1220367968 M * daniel_hozac well, yeah. just for fail over. 1220367989 M * Bertl what you can do is put the balancer on one host 1220367999 M * Bertl (declare that one the 'primary' one) 1220368017 M * Bertl and have the second one jump in (with the same ip, without balancer) if that fails 1220368048 M * Bertl and instruct the balancer on the first one not to use the 'secondary' host when it is down 1220368108 M * Bertl but that gives some kind of assymmetrical setup, unless you do this on both hosts with different ports for balancer and apache 1220368592 M * nas bert you mean let load balancer use one physical host 1220368604 M * nas sorry for late reply 1220368829 M * nas dedicated physical host for load balancer?! 1220368917 M * Bertl as I said, one option (probably the best) you can use a dedicated load balancer for that too 1220368940 M * Bertl (i.e. some piece of hardware which is designed to do load balancing) 1220369141 J * tramjoe_merin ~tramjoe@193.41.238.151 1220369298 M * ard keepalived will do the HA part for you 1220369308 M * ard in a way that's less confusing than hearbeat ;-) 1220369389 M * ard Bertl : we use dedicated linux-virtual-servers since our altheon hardware loadbalancers could not handle tcp/ip or a lot of clients 1220369459 M * ard with the right setup a lb cluster can pump more than 800Gb/s through it, and much more with direct routing 1220369474 M * ard (needs a lot of experience though ;-) ) 1220369490 M * ard anyway... /me is off to a madonna concert 1220369511 M * ard ow... s/Gb/Mb/ ;-) 1220369521 M * mnemoc =) 1220369528 M * Bertl never said that Linux as loadbalancer is bad :) 1220369552 M * daniel_hozac 800 Gbps would be pretty cool. 1220370611 M * nas thanks people 1220370735 J * _Keks_ ~bgi|id@pD9EA2C9A.dip0.t-ipconnect.de 1220370752 M * _Keks_ hi, again 1220370790 M * nas goodnight pips 1220370801 M * nas i'll think about my vserver tomorrow again 1220370842 Q * nas Quit: ChatZilla 0.9.83 [Firefox 3.0.1/2008070208] 1220370944 N * RenOS RenOS_Away 1220371764 J * balbir_ ~balbir@122.167.111.224 1220371864 J * xanoro ~xanoro@p5B169F7B.dip0.t-ipconnect.de 1220371873 Q * xanoro 1220373857 J * Bigjack ~Bigjack@85.222.113.81 1220373935 J * dna ~dna@p54BCF495.dip.t-dialin.net 1220374063 M * Bigjack Hi, I have problem with vserver one again :( While building new guest after successfullly packages download and extract I have this error message : W: Failure trying to run: chroot /etc/vservers/.defaults/vdirbase/test1 mount -t proc proc /proc . I've been searching with Google but I haven't found anything that could help with that :( Please help 1220374085 M * daniel_hozac grsec? 1220374104 M * Bigjack at the start there's a "gresec" test. 1220374109 M * Bigjack *text 1220374130 M * Bigjack how can I get rid of this ? 1220374139 M * Mr_Smoke_ grsec* 1220374144 N * Mr_Smoke_ Mr_Smoke 1220374172 M * daniel_hozac uh, don't patch the kernel with it? 1220374184 M * Bertl Bigjack: hmm, well, it is unlikely that grsec sneaked up on you :) 1220374192 M * Bigjack I want to have grsecurity in kernel :P 1220374210 M * Bigjack I have one build with vserver patch and grsecurity 1220374213 M * Bertl then disable it or configure it properly 1220374217 M * Mr_Smoke Well you can, except you might want to get rid of some of the /proc restrictions 1220374230 M * Mr_Smoke And chroot restrictions too btw 1220374246 M * daniel_hozac like the readme says. 1220374248 M * Bigjack so I just should disable /proc restrictions and chroot ? that's all ? 1220374272 M * Bigjack OK. Thanks! Going to make new kernel :P 1220374495 P * Bigjack Thanks for help! 1220374646 Q * tramjoe_merin Quit: using sirc version 2.211+KSIRC/1.3.12 1220374747 N * pmenier pmenier_off 1220374968 J * ViRUS ~mp@p579B53D6.dip.t-dialin.net 1220375817 N * DoberMann DoberMann[PullA] 1220375911 J * stragulus ~stragulus@82-197-206-246.dsl.cambrium.nl 1220375931 M * stragulus Hey 1220375960 M * stragulus Quick question... running linux-2.6.25 vserver kernel from etch-backports here, with util-vserver 0.30.214 1220375984 M * stragulus whenever I do a vserver stop, device lo on the vserver host loses its IP...is that a known prob? 1220375989 M * stragulus (and if so, what's the fix :) 1220376007 M * daniel_hozac don't assign 127.0.0.1 to guests. 1220376057 J * AlanCox ~Lee@bl7-159-43.dsl.telepac.pt 1220376177 J * dna_ ~dna@p54BCC6B5.dip.t-dialin.net 1220376235 M * stragulus Right, that does the trick. Thanks! 1220376284 M * stragulus Also, is that kernel suitable for stable use? hardware requires a more recent kernel than the 2.6.22 on the site, so I was happy to find 2.6.25-vserver in etch-backports. 1220376315 M * daniel_hozac doesn't have all the features yet, and a couple of things are known to be broken. 1220376325 M * daniel_hozac but it should work fine for most setups. 1220376424 J * vargadani ~vargadani@catv-89-135-16-217.catv.broadband.hu 1220376482 M * stragulus Ok I see, thank you. 1220376539 Q * dna Ping timeout: 480 seconds 1220376706 J * dna ~dna@p54BCC6B5.dip.t-dialin.net 1220377077 Q * dna_ Ping timeout: 480 seconds 1220377818 M * vargadani hm... remember when I was talking about ubuntu and vserver and how terrible it was to install it on ubuntu 8.04? 1220377825 M * vargadani i tried it on a server edition as well 1220377836 M * vargadani and i was surprised as there were no problems at all 1220377846 M * vargadani getting it run smoothly 1220377866 M * vargadani just wanted to tell you... 1220378037 N * DoberMann[PullA] DoberMann 1220378039 M * stragulus that's exactly what I'm using, with that kernel from etch-backports. So far, so good. 1220378055 M * vargadani i recompiled the kernel from scratch 1220378064 M * vargadani so far so good :) 1220378074 M * vargadani what version in etch running? 1220378091 M * vargadani i mean what kernel version is etch running on? 1220378107 M * stragulus 2.6.18 I think, etch-backports has 2.6.25 1220378130 M * vargadani woo 1220378134 M * vargadani i run on 2.6.22 1220378136 M * fb vargadani: 2.6.18 with tons of debian patches 1220378145 M * stragulus since ubuntu didn't seem to have a suitable vserver kernel at all, I just grabbed the deb from etch-backports and dpkg -i that 1220378162 M * vargadani is that really that simple? 1220378167 M * stragulus so far ;) 1220378172 M * stragulus surprised me too 1220378272 A * stragulus usually sticks to debian etch servers, but ubuntu server was the only os that installs without complaining on 4TB+ disks 1220378374 M * vargadani ehh 1220378386 M * daniel_hozac your system can boot off of GPT-labels? 1220378390 M * vargadani it seems that my install isn't that great 1220378399 M * stragulus it actually does, yes. 1220378402 M * vargadani vsysctl: chdir(): No such file or directory 1220378410 M * vargadani what does it mean? 1220378412 M * stragulus only caveat is that you need to reinstall grub after every partition change 1220378591 M * vargadani i cannot start my ubuntu guest 1220378594 M * vargadani i get this error 1220378613 M * vargadani and one that the error is in somewhere rc 3 1220378634 M * Bertl vsysctl sounds like util-vserver 1220378642 M * Bertl check with --debug (the startup) 1220378645 M * daniel_hozac it is. 1220378656 M * daniel_hozac that means /proc/sys isn't accessible in the guest. 1220378971 M * vargadani can i somehow make it accessible? 1220379127 M * daniel_hozac i don't know why it isn't. 1220379368 M * vargadani i just installed the one that comes with ubuntu 1220379371 M * vargadani i hope it works 1220379385 M * vargadani that is quite recent 1220379411 M * vargadani stragulus: so how does it go with etch-baclports? 1220379417 M * vargadani what does it mean at all? 1220379425 M * vargadani can you give me some info on that pls? 1220379728 M * vargadani with the util vserver that comes with ubuntu by default it works ok, however when I try to start a ubuntu guest it hangs at: Starting kernel log daemon... 1220379736 M * vargadani after pressing ctrl+C it proceeds 1220379745 M * vargadani does anybody have an idea why that could be? 1220379760 M * daniel_hozac some klogds get pissed when they can't access the kernel log. 1220379785 M * daniel_hozac just disable the initscript, or enable VXC_SYSLOG to give it a virtual, always-empty, one. 1220379799 M * vargadani could the problem be that i compiled the kernel with no debugging info? 1220379818 M * daniel_hozac i don't see how... 1220379826 M * daniel_hozac but you really want to have debugging info enabled. 1220379837 M * vargadani ok 1220379845 M * vargadani next time I will leave it there :) 1220379851 M * daniel_hozac makes it infinitely easier to track down bugs when they do happen... 1220379882 M * vargadani how do i enable that VXC_SYSLOG? 1220379905 M * daniel_hozac echo VXC_SYSLOG >> /etc/vservers//ccapabilities, as per the wiki. 1220379978 J * xdr ~xdr@gote2.16.cust.blixtvik.net 1220380022 Q * xdr 1220380152 J * xdr ~xdr@gote2.16.cust.blixtvik.net 1220380165 J * xdr_ ~xdr@gote2.16.cust.blixtvik.net 1220380170 Q * xdr_ 1220380197 Q * xdr 1220380639 J * xdr ~xdr@gote2.16.cust.blixtvik.net 1220380688 Q * xdr 1220381282 Q * virtuoso Ping timeout: 480 seconds 1220381708 M * vargadani what does the --context mean when creating a new quest? 1220381730 M * Bertl it is the unique identifier for the process context 1220381742 M * Bertl (number between 2 and 49151) 1220381938 N * RenOS_Away RenOS 1220381979 M * vargadani sweet 1220381981 M * vargadani thanx 1220382446 Q * balbir_ Ping timeout: 480 seconds 1220382983 J * ktwilight__ ~ktwilight@213.121-66-87.adsl-dyn.isp.belgacom.be 1220383068 J * xdr ~xdr@gote2.16.cust.blixtvik.net 1220383082 Q * xdr 1220383106 J * balbir_ ~balbir@122.167.66.207 1220383180 Q * ktwilight__ Remote host closed the connection 1220383216 J * ktwilight__ ~ktwilight@213.121-66-87.adsl-dyn.isp.belgacom.be 1220383237 Q * ktwilight_ Ping timeout: 480 seconds 1220383457 M * RenOS is there any way I can have the vservers have a loopback interface but not interfere with the hosts lo interface? 1220383487 M * daniel_hozac yes, use Linux-VServer 2.3+. 1220383489 M * Bertl yep, get a vs2.3.x patch and configure lback 1220383508 M * RenOS 2.6.26.2-vs2.3.0.35 1220383524 M * RenOS I'm using it already. any tips on how to configure lback? 1220383587 M * Bertl disable single ip special casing (if you only have one ip, besides 127.0.0.1) 1220383620 M * Bertl enable LBACK, unless you have it set to auto in the kernel config 1220383666 M * RenOS Bertl, ok. I'll have a check 1220383718 M * RenOS option "Automatically Assign Loopback IP" is already enabled 1220383736 M * Bertl so you are almost fine 1220383744 M * daniel_hozac and the single IP special casing? 1220383753 M * RenOS yes.. sorry to trouble you... 1220383798 M * RenOS all the guests have a loopback if! 1220383808 M * RenOS daniel_hozac, is this a kernel config option? 1220383817 M * daniel_hozac it's both. 1220383826 M * daniel_hozac the kernel has an automatic mode. 1220383918 J * xdr ~xdr@gote2.16.cust.blixtvik.net 1220384021 M * RenOS I don't have that option in the kernel config under the vserver menu 1220384048 M * daniel_hozac it should be the first option. 1220384063 M * daniel_hozac oh, no, second. 1220384134 M * RenOS v2.6.26.2-vs2.3.0.35 on this patch the second option is Enable COW Immutable Link Breaking 1220384317 M * daniel_hozac it's not. 1220384321 M * Bertl config VSERVER_AUTO_SINGLE bool "Automatic Single IP Special Casing" depends on EXPERIMENTAL 1220384340 M * daniel_hozac yeah... but building a kernel without EXPERIMENTAL is hard :-) 1220384342 M * Bertl so if you have EXPERIMENTAL off, it will be off too# 1220384351 M * RenOS oh I see 1220384451 J * Loki|muh loki@satanix.de 1220385326 Q * ktwilight__ Read error: Connection reset by peer 1220385359 J * ktwilight_ ~ktwilight@213.121-66-87.adsl-dyn.isp.belgacom.be 1220385503 Q * ktwilight_ Read error: Connection reset by peer 1220385520 J * ktwilight ~ktwilight@213.121-66-87.adsl-dyn.isp.belgacom.be 1220386027 Q * vargadani Remote host closed the connection 1220387528 J * mountie ~mountie@69-196-135-56.dsl.teksavvy.com 1220388653 J * ViRUS_ ~mp@p579B4727.dip.t-dialin.net 1220388692 J * dna_ ~dna@p54BCC6B5.dip.t-dialin.net 1220388838 J * dna__ ~dna@p54BCC6B5.dip.t-dialin.net 1220388851 Q * ViRUS_ 1220389052 Q * dna Ping timeout: 480 seconds 1220389064 J * Aiken ~Aiken@ppp118-208-124-125.lns4.bne4.internode.on.net 1220389084 Q * ViRUS Ping timeout: 480 seconds 1220389199 Q * dna_ Ping timeout: 480 seconds 1220389998 N * DoberMann DoberMann[ZZZzzz] 1220390382 Q * derjohn_foo Ping timeout: 480 seconds 1220390616 Q * RenOS Quit: later.... 1220390818 Q * _Keks_ Read error: Connection reset by peer 1220391402 J * derjohn_foo ~aj@e180205133.adsl.alicedsl.de 1220392337 Q * loddafni1 Remote host closed the connection 1220392802 J * doener_ ~doener@i577B853E.versanet.de 1220392903 Q * doener Ping timeout: 480 seconds 1220393728 J * oliwel ~chatzilla@ppp-88-217-28-254.dynamic.mnet-online.de 1220393784 M * oliwel Anybody here can help me put with some questiosn about context and mounting 1220393939 M * oliwel so - here is my problem 1220393960 M * oliwel I tried to use unionfs (aufs) together with vserver guests 1220393974 M * oliwel basic setup is a readonly root and an overlayed aufs 1220393986 M * Bertl aufs is working via fuse? 1220394002 M * oliwel Hu, no I guess not 1220394017 M * Bertl was just a question, np, please continue 1220394025 M * oliwel ok 1220394027 M * oliwel so... 1220394044 M * oliwel I have a drbd device for each overlay 1220394057 M * oliwel first, I did 1220394068 M * Bertl ah, here comes the network part ... 1220394077 M * oliwel mount the drbd device to a mountpoint in the root context 1220394091 M * oliwel put the aufs mount into the guests fstab 1220394098 M * oliwel works flawlessly 1220394101 M * oliwel now: 1220394110 M * oliwel I created a prepre-start script 1220394146 M * oliwel which sets the drbd to primary and than mounts it to an external mountpoint 1220394163 M * oliwel the aufs part is than again done with the fstab 1220394173 M * oliwel and here two things are goping wrong 1220394189 M * Bertl try with the initialize instead of the prepre :) 1220394200 M * oliwel first - I can see the mount when i type "mount", but it is "not there" - which means I see the files of the root 1220394223 M * oliwel second - when shuting down the guest - I get a nice kernel oops 1220394228 M * oliwel wtf is initalze ? 1220394243 M * Bertl kernel oops is interesting, can you upload that one? 1220394266 M * oliwel I will create a fresh one for you :) 1220394274 M * daniel_hozac use mount -n from your script. 1220394330 M * oliwel hm, moment 1220394337 M * oliwel I will create some testing setup 1220395256 M * oliwel hmm, i cant reproduce it with the test setup...I guess must kill my mailserver *duck* 1220396283 M * oliwel I guess teh drbd thing is out of scope - it looks like my mounting script is not mounting the drbd devce in place and so I just overlay two trees from the same block device, here is the kernel dump anway 1220396285 M * oliwel http://paste.linux-vserver.org/12416 1220396544 M * Bertl hmm, aufs is part of 2.6.22.19? 1220396594 M * daniel_hozac not even part of 2.6.26. :-) 1220396623 M * Bertl oliwel: what's in line 414 fs/aufs/super.c ? 1220396707 Q * dna__ Quit: Verlassend 1220396777 M * oliwel sbinfo = au_sbi(sb); 1220396841 M * oliwel AuDebugOn(!sbinfo || !list_empty(&sbinfo->si_plink)); 1220396841 M * oliwel it looks top me that the drbd mount is done somewhere in another contexct and afterwards screws up the system 1220396855 M * oliwel we had a similar issue 2 years ago when I was playing with drbd - you rememeber...? 1220396886 M * Bertl well, to me it looks like aufs is not handling namespaces properly 1220396931 M * Bertl the drbd ref counting was fixed at some point, IIRC 1220396980 M * Bertl and the line 414 is the sbinfo assignment? 1220397015 M * Bertl if so, what does au_sbi() look like? 1220397063 M * oliwel do you know how to enable line numbers in vi... 1220397078 M * oliwel this debian system has a weird vi config... 1220397092 M * Bertl http://www.mail-archive.com/aufs-users@lists.sourceforge.net/msg00834.html 1220397100 M * Bertl could that be related? 1220397204 M * oliwel I found that already 1220397212 M * Bertl it seems aufs doesn't like --bind mounts :) 1220397218 M * oliwel I use recent code, so this bug should be fixed... 1220397226 M * oliwel hm I dont make bind mounts 1220397245 M * oliwel godness - I have an idea.... 1220397247 M * oliwel mom 1220397436 M * oliwel zhm no that wasnt 1220397525 M * Bertl well, starting a guest makes at least one bind mount 1220397545 M * Bertl anyway, did you talk with the aufs developer(s)? 1220397598 M * oliwel not yet 1220397615 M * Bertl well, I guess they know the aufs code better than we do :) 1220397616 M * oliwel but as I see there is another issue which I guess is the reason 1220397623 M * oliwel sure 1220397628 M * oliwel but now - back to context 1220397632 M * Bertl okay 1220397637 M * oliwel so 1220397642 M * oliwel in prepre-start 1220397649 M * Bertl or initialize? 1220397654 M * oliwel no prepre 1220397680 M * oliwel the mounting o fthe drbd (raw device) to the mountpoint is not visibile in the guest 1220397684 M * oliwel :( 1220397707 M * oliwel any idea ? 1220397711 M * daniel_hozac and the mount point is in the guest? 1220397712 M * Bertl where do you mount what, and what do you expect to see? 1220397720 M * oliwel I guess its simply the wrong namespace 1220397723 M * oliwel ok 1220397727 M * oliwel once again 1220397748 M * daniel_hozac all the scriptlets are executed in the guest's namespace. 1220397763 M * Bertl did we see your prepre script yet? 1220397764 M * oliwel daniel_hozac: but the mount is not there 1220397775 M * daniel_hozac and what makes you say that? 1220397799 M * oliwel its just two lines 1220397833 M * oliwel # /sbin/drbdadm primary drbd1 1220397834 M * oliwel # mount /dev/drbd1 /vservers/_overlays/testing 1220397855 M * oliwel after the start the device is in primary state 1220397858 M * Bertl and your guest root is where? 1220397860 M * oliwel but it has zero links 1220397872 M * oliwel the guest root is /vservers/testing 1220397889 M * oliwel and in the guests fstab I merge the overlay to the root 1220397893 M * Bertl well /vservers/_overlays/testing is definitely outside, no? 1220397901 M * oliwel yes it is 1220397910 M * Bertl so you won't see it inside the guest 1220397910 M * oliwel but this part works if I skip the drbd stuff 1220397934 M * oliwel I will see it 1220397938 M * oliwel I do see it 1220397939 M * Bertl and most likely it gets unmounted by the cleanup 1220397950 M * Bertl where do you see it? 1220397955 M * oliwel as far as I can see, yes it does 1220397975 M * oliwel if I do the drbd stuff within the root cntext, everything is fine 1220398037 M * oliwel you talked about "initialize" - whats this ? 1220398153 M * Bertl that is the script executed outside the namespace 1220398162 M * oliwel sounds nice - is it just another directory in the scripts folder ? 1220398165 M * Bertl but I still don't see what you expect from the mount? 1220398171 M * oliwel ok 1220398179 M * Bertl I mean, it is clearly outside the guest 1220398181 M * oliwel aufs can only merge filesystems 1220398190 M * oliwel not raw devices 1220398203 M * Bertl you know, that filesystems outside are cleaned up (if not told otherwise) 1220398206 M * oliwel so I have to mount the drbd block device to a mountpoint first 1220398222 M * oliwel on exit, but not on start... 1220398232 M * Bertl yes, on startup 1220398236 M * oliwel ah ok 1220398243 M * oliwel this might be the problem than... 1220398250 M * oliwel how to "tell otherwise" 1220398283 M * Bertl namespace-cleanup-skip 1220398287 M * Bertl http://www.nongnu.org/util-vserver/doc/conf/configuration.html 1220398296 M * oliwel ok but this skips all FS... 1220398308 M * Bertl hmm? 1220398320 M * oliwel and then I guess I will recover the old issues with mounts that block my drbd devices 1220398336 M * Bertl aren't you confusing that with nonamespace-cleanup 1220398338 M * Bertl ? 1220398341 M * oliwel anyway - if I mount the filesystems in initialize - are they cleand up too ? 1220398347 M * Bertl yes 1220398381 M * Bertl that's actually the purpose of the cleanup 1220398392 M * oliwel but - if I mount my drbd OUTSIDE the guests rootfs and make the merge, it works 1220398394 M * Bertl to get rid of host specific, non guest related mounts 1220398398 M * oliwel so this is not cleaned up 1220398433 M * Bertl it should get cleaned up as well, when it is outside the guest 1220398447 M * Bertl note: I have no idea _how_ aufs handles that :=) 1220398462 M * oliwel .) 1220398491 M * Bertl but try to put the aufs template path into the namespace-cleanup-skip list 1220398509 M * Bertl and see if that fixes the issue for you 1220398543 M * Bertl (also using the -n mount option, if you insist on using the script instead of the guest's fstab is a good idea) 1220398760 M * oliwel so . if I mount the drbd device to the mountpoint from the root context, and then fire up the guest 1220398776 M * oliwel I have the mount in both contexts....so its not cleaned up anyway 1220398801 M * Bertl first, don't mix context with namespace 1220398823 M * Bertl second, I'm pretty sure the tools _try_ to unmount it 1220398832 M * oliwel :) 1220398838 M * Bertl but for some reason (aufs?) that works in some cases 1220398841 M * Bertl and fails in others# 1220398848 M * oliwel seesm that the mountpoint gtes locked 1220398863 M * oliwel the bug seems to be a counter on the drbd 1220398866 M * Bertl could be, but not necessarily at mount time 1220398997 M * Bertl maybe the lock is off by one, and having it on the host _and_ inside the guest namespace helps :) 1220399028 M * Bertl interesting would be what happens to the guest when you remove the host mount (for example) 1220399060 M * Bertl also checking that the mount is present in the guest namespace (with /proc/mounts) would be interesting 1220399126 M * oliwel tryd already- it remains in the guest 1220399226 M * oliwel ok - now it comes nice... 1220399265 M * oliwel making the drbd mount in the initialize create the mount, as the device is locked, but I am unable to see it eiterh in the guest nor via vnamespace 1220399272 M * oliwel this is all weird... 1220399448 M * Bertl please define 'see' 1220399456 M * oliwel stop - it works :) 1220399473 M * oliwel it is not in the mtab but the overlay works 1220399481 M * oliwel abd no oopses 1220399487 M * Bertl well, of course it's not in the mtab 1220399519 M * oliwel its also not in proc/mounts 1220399519 M * Bertl that's the purpose of the -n 1220399525 M * oliwel I removed the n to test 1220399537 M * Bertl well, it should better be in /proc/mounts 1220399546 M * Bertl otherwise it is not mounted :) 1220399557 M * oliwel its outside the tree... 1220399563 M * oliwel so it should not be there 1220399571 M * oliwel but it seems to behave as expected 1220399576 M * Bertl ah, it's not visible in the guest, that's fine 1220399590 M * Bertl but it _must_ be visible in the guest's namespace 1220399600 M * oliwel Noiw the question is, if we rely on an error here or on a feature 1220399605 M * Bertl i.e. check with vnamespace 1220399608 M * oliwel mom will try 1220399710 M * oliwel hmmm 1220399726 M * oliwel if I "browse" the mountpoint, its empty 1220399749 M * oliwel so it seems like the mountpoint is cleaned up but the merge of aufs remains 1220399760 M * oliwel vnamespace -e 99 ls /vservers/_overlays/testing/ 1220399764 M * oliwel reports nothing 1220399780 M * Bertl check with /proc/mounts inside the namespace 1220399798 M * oliwel but I put -n into the script 1220399817 M * Bertl -n only affects the (more or less) useless mtab 1220399855 M * oliwel its not in the mounts 1220399875 M * Bertl well, then it got officially unmounted :) 1220399889 M * oliwel http://paste.linux-vserver.org/12417 1220399892 M * oliwel *g* 1220399955 M * Bertl well, the aufs mount is there, no? 1220399962 M * Bertl what are we looking for? 1220399968 M * oliwel yes it is 1220399980 M * oliwel and if I mount the drbd device in another location 1220399990 M * oliwel I can see the changes I male inside the guest 1220399995 M * oliwel so thst seems to work