1169424984 Q * dna Quit: Verlassend 1169425513 M * micah uff, just did a vserver restart, and it kernel panic'd, although the machine is still passing traffic I can't get to it anymore, "Code: Bad EIP value. kernel: EIP: [<00000000>] _stext+0x3feffd6c/0x3c SS:ESP 0068:d0889ddc" this is 2.6.18-3-vserver-686 1169425521 M * micah the last line I saw was: /etc/rc6.d/S90reboot: line 11: 13246 Segmentation fault reboot -d -f -i 1169425659 M * micah http://paste.linux-vserver.org/912 1169425682 M * micah this vserver has bcapability CAP_SYS_RESOURCE set 1169426827 J * ridgey ~ridgey@bne-webserver.matilda.net.au 1169426911 J * facenew ~new@222.188.134.230 1169426912 M * facenew OT: a short movie mocking Kim Jong Il and his secret agent buying something from China: http://www.youtube.com/watch?v=cwppgeQQ4sA 1169427023 J * DEac- ~deac@2001:6f8:1021::20 1169427026 M * DEac- moin moin 1169427045 M * ridgey whois danielle 1169427048 M * ridgey exit 1169427075 P * facenew 1169427530 Q * ridgey Ping timeout: 480 seconds 1169429485 Q * ZLinux Ping timeout: 480 seconds 1169436244 J * Daniel15 ~Daniel15@server.daniel15.com 1169436529 M * Daniel15 Anyone know what would cause a high load average? 1169436545 M * Daniel15 I've stopped nearly everything running, and the load average is still 2.09 1169436551 M * Daniel15 http://paste.uni.cc/12806 is the top output 1169438346 N * Bertl_zZ Bertl 1169438363 M * Bertl hmm, back now ... probably off to bed again soon ... 1169438381 M * Bertl micah: you should not use those broken debian kernels :) 1169438436 M * Bertl Daniel15: what does vtop show? 1169438652 M * Daniel15 I have no virtual servers/guests running, so I think it will show the same thing 1169438670 M * Daniel15 Yeah, same thing 1169438730 M * Bertl okay, basically a load of 2 is not _that_ unusual as long as something _is_ going on 1169438759 M * Bertl although, in your case it looks like top is really the only running thing 1169438836 M * Bertl what happens when you press 'i' with top running? 1169438865 M * Bertl is top the only thing left or are there others popping up on a regular basis? 1169438911 J * ZLinux ~ZLinux@88.213.59.237 1169439000 M * Daniel15 When I press I, the only thing is top 1169439009 M * Daniel15 And it _is_ unusual for me, as it's usually 0.00 1169439023 M * Daniel15 Actually, when I press i, there's top and ssh 1169439028 M * Daniel15 sshd* 1169439030 M * Bertl okay, and the load still is at two? 1169439033 M * Daniel15 Yeah 1169439039 M * Daniel15 top - 15:07:34 up 22:59, 1 user, load average: 2.00, 2.00, 2.00 1169439050 M * Daniel15 I'll reboot it later and see if that fixes it 1169439059 M * Daniel15 I recovered from my kernel panic yesterday 1169439064 M * Daniel15 Found out it was the network card causing it 1169439066 M * Bertl okay, let's do a 'ps auxwww' right now and upload that please 1169439156 M * Daniel15 OK, http://paste.uni.cc/12808 1169439165 M * Daniel15 Is it possible this could be _another_ kernel problem? I'm still using the Debian kernel.. Should I compile my own? 1169439223 M * Bertl could be, we got so many reports of the 2.6.18-3 being broken 1169439242 M * Bertl I do not know what happened there, as none of the mainline kernels exposed any of those issues 1169439321 M * Daniel15 Should I downgrade, or compile my own? I was using 2.6.16-2-vserver-686 previously (when I wrote my tutorial), now I'm using 2.6.18-3-vserver-686 1169439348 M * Bertl I would advise to go with 2.6.19.2 and the latest 2.2.0 rc 1169439415 M * Daniel15 OK... Anything special I need to do when compiling the kernel (I've never compiled a kernel before)? 1169439455 M * Bertl probably best to use the existing debian config then, will take quite some while to compile all that stuff, but at least it should boot then :) 1169439467 M * Daniel15 OK, thanks :) 1169439469 M * Daniel15 I'll try it later 1169439479 M * Bertl np 1169439681 Q * TrueBrain Ping timeout: 480 seconds 1169439840 M * micah Bertl: :D 1169439877 M * micah any ideas what it could be? 1169439881 M * Daniel15 Well, I'll be back later... I'm going to restart the server and see if that fixes it (even temporarially) 1169439891 Q * Daniel15 Quit: Daniel15 1169440318 M * Bertl micah: probably a badly applied patch or so 1169440371 M * Bertl micah: I would start by comparing the debian kernel to mainline 1169440401 M * Bertl (and check for obvious differences) 1169441250 Q * DreamerC_ Quit: leaving 1169441266 J * DreamerC ~dreamerc@125-225-96-11.dynamic.hinet.net 1169443238 M * daniel_hozac Hollow: looks sane, but i'm not sure copying /etc/resolv.conf is something we want build to do. 1169443336 M * daniel_hozac Bertl, micah: that's the usual 2.0.2.2-rc8 bug. 1169443381 M * daniel_hozac micah: there are fixed packages at... http://kernel-archive.buildserver.net 1169443426 M * Bertl what about the load=2 issue from Daniel15? 1169443441 M * daniel_hozac i'm guessing two processes in D state? 1169443457 M * Bertl well, we could not find them ... 1169443457 M * daniel_hozac possibly from rebooting guests. 1169443461 J * TrueBrain truelight@openttd.org 1169443573 M * Bertl daniel_hozac: ah, you mean they are left over inside a context? good point! 1169443579 M * daniel_hozac they can be. 1169444096 J * ridgey ~ridgey@bne-webserver.matilda.net.au 1169444135 M * Bertl welcome ridgey! 1169444155 Q * ridgey 1169445640 M * micah daniel_hozac: are these ones you've made? 1169445707 M * daniel_hozac micah: no. they should be official kernels, if i understand things correctly. 1169445757 M * daniel_hozac but they haven't been pushed to sid because of some disagreements in the kernel team, IIRC. 1169446262 M * micah daniel_hozac: ah, I haven't been able to follow the last month, thanks 1169446322 M * daniel_hozac np. 1169446366 M * Bertl do we have them on a wiki page, btw? 1169446501 M * Bertl ala 'Known Good Kernels' or so? 1169446561 J * Ecelimsin Sohbet@88.230.247.69 1169446570 M * Bertl welcome Ecelimsin! 1169446586 M * Ecelimsin welcome Bertl 1169446633 J * sladen paul@starsky.19inch.net 1169446673 P * Ecelimsin 1169447319 J * sladen_ paul@starsky.19inch.net 1169447319 Q * sladen Read error: Connection reset by peer 1169448193 Q * ntrs_ Ping timeout: 480 seconds 1169448432 J * gab ~gab@158.36.45.236 1169448736 M * Bertl welcome gab! 1169449060 Q * comfrey Ping timeout: 480 seconds 1169449905 J * Fire_Egl ~FireEgl@adsl-63-199-170.bhm.bellsouth.net 1169450300 Q * FireEgl Ping timeout: 480 seconds 1169450534 M * DEac- my nic has now 3 ips. 2 for my root system, 1 for my virtual system. but on all 3 ips only pids of the root system listens. how i change it, that the virtual system get one ip? 1169450568 M * Bertl hmm, how do you verify that? 1169450587 M * DEac- Bertl: with nmap 1169450597 M * DEac- and i tried to start sshd 1169450602 M * Bertl ah, from outside, yes? 1169450645 M * DEac- sshd said, somebody already listens on his port 1169450651 M * DEac- and inside 1169450718 M * Bertl the 'guest' ip shows up on the host with 'ip addr ls'? 1169450724 M * DEac- yes 1169450747 M * Bertl does 'ping -I www.google.com' work? 1169450774 M * DEac- yes 1169450808 M * Bertl does 'lsof -i :22' on the host list * or 0.0.0.0 ? 1169450845 M * DEac- in virtual? i must install it 1169450851 M * Bertl nope, on the host again 1169450867 M * DEac- ah, ok, then i must try the ping again 1169450903 M * DEac- ok, yes i can ping from host 1169450913 M * Bertl and the lsof part? 1169450939 M * DEac- sshd listens on it 1169450952 M * Bertl like this: *:ssh (LISTEN) ? 1169450959 M * DEac- yes 1169450970 M * Bertl so, that is the problem why your guest sshd won't start 1169450985 M * Bertl the host is not restricted in binding like the guests 1169450994 M * DEac- ok, then i must change it in host, that it doesn't use this ip? 1169450997 M * Bertl so an sshd on the host will bind _all_ ips 1169451015 M * Bertl yep, just add a Listen directive specifying the host only ips 1169451047 M * DEac- it's possible to restrict it on host, that processes can listen on * without this ip? 1169451050 M * Bertl /etc/ssh/sshd_config ListenAddress a.b.c.d:22 1169451081 M * Bertl yes, but you do not want that for sshd, unless you do not plan to administrate the machine via sshd :) 1169451114 M * Bertl because if you restrict the host to a certain IP subset, you will not be able to start guests on a different ip set :) 1169451186 M * DEac- what's the meaning of 'ip set'? 1169451198 M * Bertl a bunch of ip addresses 1169451224 M * DEac- ah, so, if i restrict my host, i restict my virtual systems too 1169451247 M * Bertl if you ssh into that 'restricted' sshd, yes 1169451304 M * DEac- i mean, not only the sshd, i mean all processes, which listen on * shouldn't list on the 'virtual ip' 1169451325 M * Bertl well, you can 'jail' the init, which will do that 1169451347 M * Bertl but you will not be able to use any ip outside that scope on the host anylonger, which kind of makes it useless 1169451395 M * DEac- that's bad 1169451399 M * Bertl why? 1169451426 M * Bertl typically you do not run stuff on the host, except for sshd, because there is no real reason for doing so 1169451427 M * DEac- because i must change many daemons 1169451444 M * DEac- ok, that's true 1169451447 M * Bertl move those daemons to a guest 1169451503 M * DEac- it's possible to use ipv6? 1169451518 M * Bertl yes, with recent kernels and a patch 1169451578 M * DEac- ok, i found it, thanks 1169452161 Q * Aiken Quit: Leaving 1169453246 J * duckx ~Duck@tox.dyndns.org 1169453880 M * matti Bertl: :) 1169453911 M * Bertl hey matti! 1169453996 M * matti Morning Bertl :) 1169454691 J * FireEgl ~FireEgl@adsl-147-90-23.bhm.bellsouth.net 1169455070 Q * Fire_Egl Ping timeout: 480 seconds 1169455130 M * DEac- how i change which programm would be started as init? 1169455272 M * Bertl you know the wiki page describing the config tree? 1169455319 M * DEac- something like that, i search 1169455329 M * Bertl http://linux-vserver.org/util-vserver:Documentation 1169455342 M * DEac- ah, thanks 1169455354 M * Bertl np 1169455446 Q * shedi Quit: Leaving 1169455755 Q * renihs Quit: Leaving 1169455841 J * ntrs ~ntrs@68-188-55-120.dhcp.stls.mo.charter.com 1169456337 J * renihs ~penguin@83-65-34-34.arsenal.xdsl-line.inode.at 1169456947 M * Bertl okay, off to bed again ... cya later ... 1169456951 N * Bertl Bertl_zZ 1169456959 J * dna ~naucki@152-208-dsl.kielnet.net 1169457305 M * harry cya Bertl_zZ 1169458155 J * _cob ~cob@pc-csa01.science.uva.nl 1169458282 Q * ||Cobra|| Ping timeout: 480 seconds 1169458335 J * cdrx ~legoater@cap31-3-82-227-199-249.fbx.proxad.net 1169460243 J * dlezcano ~dlezcano@blueice3n1.uk.ibm.com 1169460269 M * dlezcano Hi all 1169460876 J * toom ~toom@coco.docisland.org 1169460904 M * toom hi 1169460983 M * toom I've just installed 2.6.19.2-grsec2.1.10-vs2.2.0-rc8 and I get "ncontext: vc_net_create(): Invalid argument" when I try to start the vserver 1169461002 M * toom this is a known problem ? 1169461143 Q * ZLinux Quit: Leaving 1169461167 J * ZLinux ~ZLinux@88.213.59.237 1169461687 Q * DavidS Quit: Leaving. 1169462174 J * shedi ~siggi@dsl-149-109-85.hive.is 1169463832 M * harry toom: not that i know of.. 1169463928 M * toom It seems that this is link to the context number 1169463958 M * toom but chcontext command works 1169464294 M * harry does it work with a non-grsec kernel? 1169464305 M * harry (that eliminates the grsec being the problem ;)) 1169464311 M * toom I don't know 1169464329 M * harry hmm... i doubt its grsec related, but you can never be sure ;) 1169464338 M * toom I can't now 1169464350 M * harry i've never seen that error, so i'm afraid i can't help you 1169464355 M * toom I try to remove grsec tomorow 1169464362 M * toom ok 1169464389 M * harry just for testing... :) 1169464409 M * harry bertl, daniel_hozac etc... will definately know what this is about 1169464449 M * harry http://paste.linux-vserver.org/784 1169464452 M * harry is it something like this? 1169464473 M * toom yes exactly 1169464521 M * harry what version of utils do you have? 1169464527 M * harry 0.30.212? 1169464531 M * toom yes 1169464545 M * harry is it a new install? 1169464551 M * harry or did it work with other kernel/tools? 1169464576 M * harry do you use dynamic context id's? 1169464577 M * toom I worked with another kernel few weeks ago 1169464587 M * toom I don't know 1169464622 M * toom but I've just create a new vserver 1169464636 M * harry what does cat /etc/vservers//context say? 1169464865 M * harry ? 1169465880 Q * tamitall Ping timeout: 480 seconds 1169466404 Q * bXi Remote host closed the connection 1169467435 Q * ensc Killed (NickServ (GHOST command used by ensc_)) 1169467445 J * ensc ~irc-ensc@p54B4DBCD.dip.t-dialin.net 1169467448 P * m4z http://www.catb.org/~esr/faqs/smart-questions.html 1169468833 M * sid3windr vhoststealer! 1169469419 M * Ramjar i have a directory i want to share to the rest of vserver hosts, so i use fstab in order to mount this directory with: /folder /folder none bind,rw 0 0 i fstab. Only problem is that i only want root to have access to this folder. What can i do, in order to get this work? 1169469750 Q * softi42 Ping timeout: 480 seconds 1169469819 M * Ramjar problem solved. 1169470353 J * softi42 ~softi@p549D73B4.dip.t-dialin.net 1169470950 M * toom harry: excuse me, I had to go out 1169470973 M * toom harry: I have no context file in /etc/vservers/ 1169471060 M * toom when I create one, I get "can not determine xid of vserver ''; returned value was ''" 1169471782 M * harry you should use static contexts 1169471788 M * harry otherwise things will break... 1169472040 M * toom how can I do that ? Where I can find a doc ? 1169472114 M * harry http://linux-vserver.org/util-vserver:Documentation 1169472118 M * harry search for context 1169472152 M * toom thank you 1169472154 M * harry http://linux-vserver.org/Frequently_Asked_Questions#I_now_get_errors_like_.22ncontext:_vc_net_create.28.29:_Invalid_argument.3B_dynamic_contexts_disabled..22_on_startup._Vservers_are_not_started 1169472159 M * harry for your specific question 1169472588 J * Piet hiddenserv@tor.noreply.org 1169472602 M * harry toom: so check the faq section on the site, your problem is explained there... 1169472744 M * toom harry: I've created /etc/vservers//context but when I start my vserver I get "vshelper.init: can not determine xid of vserver ''; returned value was ''" 1169472877 M * harry toom: what does the context file contain??? 1169472917 M * harry what number did you put in the file? 1169472921 M * harry I now get errors like "ncontext: vc_net_create(): Invalid argument; dynamic contexts disabled." on startup. Vservers are not started 1169472925 M * harry Dynamic context are disabled by default and are deprecated. For example, tagxid and network checks won't be useable with dynamic ids. Now you should manually assign a explicit context to your vservers, like echo 101 > /etc/vservers/myvserv/context 1169472936 M * toom I tried 10 49152 and 101 1169472970 M * harry don't tell me your virtual server is called: 1169472981 M * toom no :) 1169472988 M * harry 14:32 < toom> harry: I've created /etc/vservers//context but when I start my vserver I get "vshelper.init: can not determine xid of vserver ''; returned value was ''" 1169473002 M * harry then why does it say: vserver '' ? 1169473004 M * toom s//atlas/g 1169473013 M * harry ah 1169473085 M * toom do I need to remove all files in /var/run/vservers.rev/ ? 1169473118 M * toom there are link like -> /etc/vservers/ 1169473195 M * harry don't know actually 1169473303 M * harry when you start your server now, what exactly is the error message? 1169473306 M * toom I tried to remove it but it doesn't solve my pb 1169474225 J * tamitall ~tam@gw.nettam.com 1169474451 M * toom I create a new vserver specifying context (newvserver --context) and there is the same pb, after unpacking the packages I get 1169474454 M * toom vshelper.init: can not determine xid of vserver 'atlas2'; returned value was '' 1169474661 M * harry odd 1169474885 M * harry http://www.paul.sladen.org/vserver/archives/200609/0230.html 1169474927 J * yarihm ~yarihm@hg-public-dock-124-dhcp.ethz.ch 1169475079 M * toom It works ! thank a lot 1169475174 M * harry np 1169476432 Q * cdrx Ping timeout: 480 seconds 1169477323 J * jabra ~jabra@70.90.101.105 1169477328 M * jabra morning 1169477361 J * lilalinux ~plasma@dslb-084-058-204-152.pools.arcor-ip.net 1169478441 J * DavidS ~david@chello062178045213.16.11.tuwien.teleweb.at 1169478619 J * cdrx ~legoater@blueice1n1.uk.ibm.com 1169479109 Q * FireEgl Quit: ... 1169480574 Q * DavidS Quit: Leaving. 1169480699 P * toom 1169480871 Q * Loki|muh Remote host closed the connection 1169480874 J * Loki|muh loki@satanix.de 1169481011 J * thunder18 ~thu@tor-irc.dnsbl.oftc.net 1169481067 Q * thunder1 Ping timeout: 480 seconds 1169481085 N * thunder18 thunder1 1169481217 Q * yarihm Quit: Leaving 1169481406 Q * Loki|muh Remote host closed the connection 1169481441 Q * gab Remote host closed the connection 1169481463 Q * kugg Ping timeout: 480 seconds 1169481896 J * Loki|muh loki@satanix.de 1169482032 J * kugg ~kugg@90-227-91-196-no120.tbcn.telia.com 1169484833 J * stefani ~stefani@tsipoor.banerian.org 1169485896 Q * Johnnie Ping timeout: 480 seconds 1169485973 J * bonbons ~bonbons@83.222.37.103 1169485984 J * Johnnie ~jdlewis@66.199.231.180 1169486901 Q * dlezcano Read error: Connection reset by peer 1169487860 N * Bertl_zZ Bertl 1169487865 M * Bertl morning folks! 1169488173 P * Johnnie G'bye! 1169488184 J * Johnnie ~jdlewis@66.199.231.180 1169488268 Q * mnemoc Ping timeout: 480 seconds 1169488642 J * mnemoc ~amery@kilo105.server4you.de 1169489139 Q * cdrx Read error: Connection reset by peer 1169489426 M * jabra Bertl: hey 1169489428 M * jabra you around? 1169489432 M * Bertl yep 1169489468 M * jabra so after i rsynced the /dev from the running vserver to the copy from backupninja it worked 1169489480 M * jabra problem is everyone is owned by the backupninja user 1169489489 M * Bertl okay, so you probably just need a few options to that 1169489504 M * Bertl i.e. save user/owner/permissions/devices/etc 1169489524 M * jabra right, wondering how you do backups for backupninja 1169489535 Q * duckx Quit: Client exiting 1169489537 M * jabra thought you had suggested backupninja in the first place 1169489549 M * jabra using rdiff for the comparisons 1169489553 J * duckx ~Duck@tox.dyndns.org 1169489726 M * Bertl jabra: nope, never used backupninja 1169489732 M * jabra ugh 1169489734 M * jabra aight 1169489738 M * Bertl I suggested dump/restore for ext2/3 1169489742 M * jabra ok 1169489750 M * Bertl and rsync for moving stuff around 1169489753 M * jabra where were the flags I needed again 1169489758 M * jabra for dump/restore ? 1169489769 M * Bertl you can do a dump like this: 1169489790 M * Bertl dump 0zf /path/to/data.dump /path/to/guest 1169489798 M * Bertl and restore it again, like this: 1169489817 M * Bertl (cd /path/to/guest; restore rf /path/to/data.dump) 1169489904 M * jabra so i could do a dump then transfer that 1169490003 M * Bertl yes, you can also do that in one line without storing the dump 1169490011 M * Bertl that would look like this: 1169490034 M * Bertl dump 0zf - /path/to/guest | ssh root@destination "(cd /path/to/guest; restore rf -)" 1169490043 M * Bertl or this: 1169490065 M * Bertl ssh root@source "dump 0zf - /path/to/guest" | (cd /path/to/guest; restore rf -) 1169490073 M * jabra thinking i script it to just dumpt he results to a file 1169490098 J * prae ~benjamin@foxhound.sherpadown.net 1169490160 J * Piet_ hiddenserv@tor.noreply.org 1169490166 Q * shedi Quit: Leaving 1169490199 Q * Roey Ping timeout: 480 seconds 1169490565 Q * Piet Ping timeout: 480 seconds 1169490970 M * jabra Bertl: what about : dump 0zf -f root@dest:/dump/server.dump /path/host 1169491055 M * Bertl I think that will give you strange results at best 1169491062 M * jabra why? 1169491096 M * Bertl the 'f' in 0zf already says output to the following argument 1169491096 M * jabra the idea being i want to have a file that i can revert to for a given day 1169491106 M * Bertl which will be -f in your case 1169491118 M * jabra ok i mean without the -f 1169491136 M * jabra dump 0zf root@dest:/dump/server.dump /path/host 1169491147 M * Bertl I don't think the dump will handle the root@dest without a tape server on the other end 1169491153 M * Bertl but you can try 1169491176 M * jabra it should in my opinion 1169491185 M * jabra will need to test it 1169491208 M * Bertl the ssh, btw can be done in a fully automated way 1169491233 M * Bertl another option for pure daily backups would be: 1169491243 M * Bertl to use rsync to 'just' sync the changes 1169491261 M * Bertl and then, on the backup machine, usedump/rdiff to make the 'snapshot' 1169491273 M * jabra well i want the thing automated like backupninja 1169491286 M * jabra so all i need to do is transfer off the backup server to do DR 1169492641 J * sharkjaw ~gab@216-160-24.0503.adsl.tele2.no 1169492768 Q * lilalinux Quit: Leaving 1169493152 Q * Greek0 Remote host closed the connection 1169493244 J * shedi ~siggi@ftth-237-144.hive.is 1169493453 J * _dmax ~semaj@81.193.56.29 1169493796 Q * dmax Ping timeout: 480 seconds 1169493799 N * _dmax dmax 1169493908 J * Greek0 ~greek0@85.255.145.201 1169494010 J * Roey ~katz@h-69-3-4-130.mclnva23.covad.net 1169495405 Q * softi42 Ping timeout: 480 seconds 1169497052 J * DavidS ~david@217.19.46.18 1169497249 Q * morfoh Ping timeout: 480 seconds 1169497252 Q * mnemoc Ping timeout: 480 seconds 1169497642 M * Bertl okay, off for now ... back later ... 1169497648 N * Bertl Bertl_oO 1169498219 J * dlezcano ~dlezcano@AToulouse-252-1-104-174.w86-213.abo.wanadoo.fr 1169499367 J * cdrx ~legoater@cap31-3-82-227-199-249.fbx.proxad.net 1169500131 J * meandtheshel1 ~markus@85-124-39-225.dynamic.xdsl-line.inode.at 1169500794 J * yarihm ~yarihm@84-75-123-221.dclient.hispeed.ch 1169500844 Q * prae Read error: Connection reset by peer 1169500943 Q * sharkjaw Remote host closed the connection 1169502148 Q * bonbons Quit: Leaving 1169502442 J * Piet__ hiddenserv@tor.noreply.org 1169502830 Q * Piet_ Ping timeout: 480 seconds 1169504018 N * Piet__ Piet 1169504214 P * click [IRSSI] 1169504580 Q * yarihm Quit: Leaving 1169504679 J * click click@ti511110a080-3802.bb.online.no 1169504730 Q * TrueBrain Ping timeout: 480 seconds 1169505950 Q * DavidS Ping timeout: 480 seconds 1169506279 Q * dna Quit: Verlassend 1169506476 J * PowerKe ~tom@d54C13E4B.access.telenet.be 1169506722 Q * jabra Ping timeout: 480 seconds 1169506750 M * PowerKe I'm trying to change files in the vservers from the host system, but I'm getting touch: cannot touch `/vservers/management/usr/share/man/man1/cat.1.gz': Permission denied 1169506765 M * PowerKe I'm guessing it has something to do with the UI flags when using showattr 1169506788 M * PowerKe But I don't understand it entirely, nor do I understand why some files get UI and some get ui flags... 1169506826 M * PowerKe It seems like I never ran into this problem on my previous vserver (which hasn't been updated for a while) 1169506833 M * Bertl_oO PowerKe: files with UI are 'shared' 1169506855 M * Bertl_oO it means that they are immutable but 'unlink'-able 1169506855 M * PowerKe They get changes from ui to UI when running vhashify? 1169506864 M * PowerKe *changed 1169506872 M * Bertl_oO yep, but that doesn't prevent accessing them 1169506890 M * Bertl_oO what might keep you from accessing them is if they have wrong tagging 1169506914 M * Bertl_oO but more likely the directory permissions (up to guest root) are in your way 1169506951 M * PowerKe I put setattr --barrier /vservers 1169506970 M * PowerKe laptoptom ~ # showattr /vservers 1169506970 M * PowerKe ---Bui- /vservers 1169506970 M * PowerKe ---bui- /vservers/management 1169506970 M * PowerKe ---bui- /vservers/sandbox 1169506992 M * Bertl_oO check the normal permissions too 1169507011 M * Bertl_oO also check the dmesg/kernel log for warnings 1169507012 M * PowerKe vservers is 000, subdirs are 755 1169507062 M * Bertl_oO inside the guest, what does the following show: 1169507071 M * Bertl_oO ls -lad / /.. 1169507091 M * PowerKe laptoptom ~ # vserver management enter 1169507091 M * PowerKe management / # ls -lad / /.. 1169507091 M * PowerKe drwxr-xr-x 17 root root 4096 Jan 20 18:06 / 1169507092 M * PowerKe drwxr-xr-x 17 root root 4096 Jan 20 18:06 /.. 1169507107 M * Bertl_oO okay, that looks fine 1169507138 M * PowerKe filesystem is jfs (if that helps anything) 1169507158 M * PowerKe mounted rw,noatime 1169507178 M * Bertl_oO are you trying to access the files as root or as user? 1169507200 M * PowerKe I'm trying to access the files as root 1169507215 M * PowerKe from the host 1169507225 M * Bertl_oO cat does work fine? i.e. read access? 1169507231 M * PowerKe yes 1169507251 M * Bertl_oO what about touching a new file in the same dir? 1169507266 M * PowerKe that works as well 1169507285 M * Bertl_oO what does lsxid report on those files (on the host)? 1169507307 M * PowerKe (btw, I get the same error on files with ui (lowercase)) 1169507328 M * Bertl_oO yeah, I think that is not related at all 1169507330 M * PowerKe laptoptom ~ # lsxid /vservers/management/usr/share/man/man1/perltoot.1.gz 1169507330 M * PowerKe !!ERR!! /vservers/management/usr/share/man/man1/perltoot.1.gz 1169507359 M * Bertl_oO okay, what kernel/patch do you use? 1169507375 M * PowerKe 2.6.17-vs2.0.2.1-gentoo 1169507396 M * Bertl_oO would it be possible to try with a more recent version? 1169507424 M * PowerKe I think that's the most recently version marked as stable, but I could switch to ~x86 I guess 1169507431 M * Bertl_oO (something like 2.6.19.2 + 2.2.0-rc*) 1169507541 M * PowerKe seems like I need to sync my portage tree first, but would 1169507578 M * PowerKe seems like I need to sync my portage tree first, but would vserver-sources-2.2.0_rc6 be good? 1169507609 M * Bertl_oO no guarantees, but I remember some jfs issues a while back, cannot verify it atm though 1169507627 M * PowerKe Ok, I'll sync, recompile the kernel and see what happens 1169507697 M * PowerKe I think I'm trying to do the same things like on my other host running 2.6.14-vs2.0.1-gentoo which doesn't show any problems (also on jfs). Maybe something broke in a newer version. 1169507761 M * Bertl_oO you could try to rsync the guest over to the other host 1169507857 M * PowerKe I'm also not having the problem on all files 1169507885 M * PowerKe What I'm trying to do is changing ctimes on different vguests so that they match so they can be vunified 1169507915 M * Bertl_oO vunify should not care about the times 1169507916 M * PowerKe The script always ran fine on the old host, but now I'm getting a lot of complaints from files that can't be changed 1169507949 M * Bertl_oO i.e. it unifies based on the checksum 1169507964 M * PowerKe Yes, but you can't have 2 files with the same inode and different ctime 1169507985 M * PowerKe I think it matches on checksum, but also verifies the other properties are the same 1169508009 M * PowerKe At least I don't get it to work without first giving both files the same ctime (unless I'm using a version that's too old) 1169508030 M * Bertl_oO you might want to talk to daniel_hozac, but I'm pretty sure the time is not relevant 1169508047 M * Bertl_oO uid/gid and modes are of relevance though 1169508180 M * PowerKe Those are the same. Both gentoo guests 1169508196 M * PowerKe But even when using binary packages, it always assigns ctime when emerging 1169508229 M * Bertl_oO i.c. yes, that actually might need special handling ... 1169508279 M * Bertl_oO what about changing the ctime on the host? 1169508287 M * Bertl_oO or is that already what you are doing? 1169508302 M * PowerKe So on my other vserver I have a script that takes the lowest ctime for all guests and then assigns that ctime to identical files in all guest whilst also updating the portage database. 1169508308 M * PowerKe I'm running the script from the host 1169508322 M * Bertl_oO hmm, and you get the permission denied there? 1169508331 M * PowerKe on the other host it has always worked perfectly 1169508344 M * PowerKe But on my laptop where I recently installed vserver it seems to fail 1169508380 M * PowerKe 2.6.14-vs2.0.1-gentoo (working) vs 2.6.17-vs2.0.2.1-gentoo (failing) 1169508467 M * Bertl_oO double check the jfs mount options, aside of that, the kernel upgrade is all I can suggest right now 1169508493 M * PowerKe Ok, I might have found something: on the 2.6.14 I enabled "JFS POSIX Access Control Lists" in the kernel 1169508610 M * PowerKe mount options are the same on both hosts (rw,noatime) 1169508648 M * Bertl_oO check with /proc/mounts, not 'mount' 1169508722 M * PowerKe working: /dev/vg0/data /data jfs rw,noatime,integrity 0 0 1169508726 M * PowerKe failing: /dev/root / jfs rw,noatime 0 0 1169508746 M * Bertl_oO integrity? 1169508788 M * PowerKe No idea where that comes from (unless it's the "JFS POSIX Access Control Lists" or "jfs statistics" in the kernel) 1169508793 M * PowerKe This is what's in fstab: /dev/vg0/data /data jfs noatime 0 2 1169508858 M * PowerKe I just recompiled the kernel on my laptop with "JFS POSIX Access Control Lists" so I can reboot and see if that changes something? 1169508871 M * Bertl_oO okay 1169508883 M * PowerKe be back in a few minutes 1169508890 Q * PowerKe Quit: Rebooting 1169509170 J * PowerKe ~tom@d54C13E4B.access.telenet.be 1169509188 M * PowerKe At least it doesn't add the integrity... 1169509191 M * PowerKe /dev/root / jfs rw,noatime 0 0 1169509227 M * PowerKe and lsxid still returns !!ERR!! 1169509265 M * PowerKe The only difference in jfs kernel options now is "jfs statistics" 1169509272 M * Bertl_oO yes, that's fine without tagging enabled 1169509284 M * Bertl_oO (the lsxid -> err) 1169509333 M * PowerKe touch still gives permission denied 1169509349 M * Bertl_oO okay, did you check dmesg? 1169509369 M * Bertl_oO next step is to run strace -fF -o touch.trace touch ... 1169509517 M * PowerKe seems like I don't have that yet, emerging... 1169509544 M * PowerKe btw, if it's too much trouble, I can just try to upgrade to a higher version 1169509559 M * Bertl_oO nah, let's check it 1169509580 M * PowerKe Now I'm hoping I didn't do something really stupid :) 1169509620 M * PowerKe laptoptom ~ # strace -fF -o touch.trace touch -r /vservers/management/usr/share/man/man1/perltoot.1.gz /vservers/management/usr/share/man/man1/perltoot.1.gz 1169509624 M * PowerKe touch: cannot touch `/vservers/management/usr/share/man/man1/perltoot.1.gz': Permission denied 1169509675 M * Bertl_oO hmm, that is the same file as reference, no? 1169509678 M * PowerKe yes 1169509697 M * PowerKe laptoptom ~ # strace -fF -o touch.trace touch 1169509697 M * PowerKe touch: cannot touch `/vservers/management/usr/share/man/man1/perltoot.1.gz': Permission denied 1169509699 Q * duckx Remote host closed the connection 1169509702 M * PowerKe without the reference 1169509725 M * Bertl_oO okay, we want the trace outputs, please upload somewhere 1169509878 M * PowerKe http://powerpnt.dyndns.org:8080/vserver/ 1169509997 M * PowerKe I didn't delete the file between running with and without running, so I'm not sure if the second run appended or wrote over the file (but I guess you know) 1169510015 M * Bertl_oO that's fine 1169510045 M * PowerKe Since there's only 1 Permission denied I'm guessing it truncated the file 1169510069 P * stefani I'm Parting (the water) 1169510189 J * Aiken ~james@tooax6-088.dialup.optusnet.com.au 1169510376 M * PowerKe integrity seems to be a default flag for jfs: Default. Commit metadata changes to the journal. Use this option to remount a volume where the nointegrity option was previously specified in order to restore normal behavior.