1255565109 J * riotfish ~riotfish@port160.ds1-ynoe.adsl.cybercity.dk 1255565140 M * riotfish hi 1255565169 M * riotfish i was wondering if the ubuntu jaunty kernel is still considered experimental like it says in http://linux-vserver.org/Installation_on_Ubuntu ? 1255565221 J * halfd ~halfd@78.129.201.121 1255566615 M * Bertl riotfish: no idea .. maybe the ubuntu folks know? 1255566643 M * riotfish Bertl: well, i've decided to compile my own kernel :) 1255566647 M * Bertl in any case, you can compile your own (mainline) kernel with Linux-VServer patches and should be fine 1255566650 M * riotfish but thnx anyways :) 1255566654 M * Bertl np 1255566656 M * riotfish heh heh 1255566684 M * Bertl 2.6.31.x? 1255566697 M * riotfish nope 1255566709 M * riotfish 2.6.22.19 1255566720 M * riotfish following http://linux-vserver.org/Installation_on_Linux_2.6 1255566747 M * daniel_hozac you're supposed to replace the version numbers. 1255566747 M * daniel_hozac :) 1255566801 M * Bertl maybe 2.6.22.19 is the version of choice :) 1255566807 M * riotfish daniel_hozac: yeah i know, but the frontpage of linux-vserver.org says that 2.6.22.19 is the most stable 1255566813 M * riotfish or that is how i read it 1255566836 M * Bertl well, it is the latest stable release so far :) 1255566851 M * riotfish i just want to be sure not to end in any unstable version-sh*t :) tried that to many times :) 1255566868 M * riotfish and since i'm doing it for a production server it seems as the best approach ;) 1255566880 M * mugwump Well, 2.6.27.35-vs2.3.0.6 has crashed on me strangely, I can't recommend that 1255566898 M * riotfish see ;) 1255566909 M * daniel_hozac well... 1255566923 M * daniel_hozac you do realize 2.6.22 has a lot of unpatched bugs, right? 1255566925 M * mugwump probably lxc is the way forward for 2.6.29 on, if it does all you need 1255566952 M * riotfish daniel_hozac: hmm... yeah good point 1255567014 M * riotfish daniel_hozac: which version would you recommend with? (my base OS is ubuntu jaunty 64bit with a 2.6.28-15-server kernel) 1255567028 M * daniel_hozac Bertl has a better idea there. 1255567094 M * Bertl mugwump: did you use/try lxc on 2.6.29+? 1255567105 M * riotfish daniel_hozac: so, just to be sure (don't have much experience with kernels), there is no problem in using the newest stable kernel from kernel.org and my current OS's kernel? 1255567130 M * mugwump Bertl: we've been experimenting with it here at Catalyst 1255567130 M * daniel_hozac the kernel is massively backwards compatible. 1255567139 M * daniel_hozac user-interfacing APIs do not change. 1255567148 M * Bertl mugwump: and, got any guest running already? 1255567149 M * riotfish k... then i'll try with 2.6.31.x 1255567248 M * riotfish daniel_hozac: what is the worst i can expect by using 2.6.31.4 and vs2.3.0.36.18 ? 1255567291 M * mugwump Bertl: a colleague reported some success, yes 1255567293 M * daniel_hozac there is a small possibility of it eating your babies. 1255567296 M * mugwump we're just tinkering so far 1255567327 M * Bertl well, let me know when it actually works then :) 1255567350 M * daniel_hozac and when it's fully functional. 1255567362 M * daniel_hozac last time i really looked at it, kill -1 in a guest would kill every process on the host. 1255567375 M * daniel_hozac and patches to fix the problem weren't merged. 1255567377 M * riotfish daniel_hozac: heh heh very funny :) but seriously? and does vs2.3.0.36.18 work with 2.6.31.4 ? (it says 2.6.31.3 on the download page) 1255567407 M * Bertl riotfish: worst case scenario is that your kernel crashes, you record a stack trace/dump via serial console (or similar) and report it to us, we'll fix it then ASAP ... (basically the same as with vanilla stable kernels) 1255567432 M * riotfish Bertl: k, but do the versions i've mentioned above work with each other ? 1255567474 M * Bertl kernels do not work with eachother ... there is only one running (highlander style :) 1255567511 M * Bertl but you can use recent kernels with userspace used to 2.4 series kernels without any problems, if that is what you are asking :) 1255567514 M * riotfish Bertl: heh heh, i'm talking about vs2.3.0.36.18 and kernel 2.6.31.4 1255567539 M * Bertl give me a minute to check that :) 1255567548 M * riotfish Bertl: great, thnx :) 1255567787 J * riotfish_ ~riotfish@port160.ds1-ynoe.adsl.cybercity.dk 1255567788 Q * riotfish Read error: Connection reset by peer 1255567793 N * riotfish_ ritofish 1255567796 M * ritofish damn... 1255567837 M * mugwump daniel_hozac: well, it's easy to snipe from the sidelines, but fortunately lxc have a modern development process, so I can quickly search the git log for that bug 1255567838 M * ritofish i accidentally turned of my computer by pressing my foot against the power supply button 1255567846 M * mugwump and see that it was fixed in 2.6.28 1255567853 M * ritofish Bertl: what's your conclusion on the versions ? 1255567909 M * Bertl good and bad news ... 1255567916 M * mugwump but I see that there was a 3-4 month window where the bug was reported and there was no stable release to fix it 1255567933 M * Bertl ritofish: the patch for 2.6.31.3 won't apply to 2.6.31.4 1255567948 M * daniel_hozac mugwump: no, it was present for, umm, releases. 1255567965 M * Bertl ritofish: the good news is, there will be a patch for 2.6.31.4 pretty soon :) 1255567979 M * ritofish what's the ETA for that? 1255567995 M * Bertl about 5 minutes, I'd say ... 1255567999 M * ritofish hah hah 1255568002 M * ritofish awesome 1255568034 M * ritofish then my plan about getting this working today isn't that unlikely :) 1255568132 M * Bertl very likely :) 1255568146 M * mugwump daniel_hozac: 3 months covers a lot of releases with linux doesn't it 1255568222 M * daniel_hozac mugwump: not really, that's about one cycle. 1255568223 M * Bertl mugwump: (snipering back) maybe Linus will eventually improve git in such way, that it fixes reported issues automatically :) 1255568314 M * mugwump what I mean is that I can date the time that your comment was accurate to between 12 and 15 months ago in seconds using the commit log 1255568336 M * mugwump actually it took a few minutes because I needed to 'git fetch' to update my clone 1255568352 M * Bertl you have a clone now? :) 1255568381 M * Bertl ritofish: kernel is test-compiling ... 1255568387 M * ritofish nice 1255568403 N * ritofish riotfish 1255568629 J * geb ~geb@earth.gebura.eu.org 1255568791 M * Bertl riotfish: test booting now :) 1255568811 M * riotfish i can't wait! :D 1255568893 M * Bertl works fine, uploading patch ... 1255568897 M * riotfish uhhh 1255568939 M * Bertl so actually it took about 15 minutes, but that's just because my buildhost is a little slow :) 1255568947 M * riotfish heh heh no big deal :) 1255568955 M * Bertl patch-2.6.31.4-vs2.3.0.36.19.diff (usual place) 1255569003 M * riotfish got it :) 1255569123 M * Bertl have fun! 1255569158 M * riotfish Bertl: thnx :) this is probably not the last you've heard from me ;) 1255569243 M * Bertl good, feel free to hang around :) 1255569362 M * riotfish i will :) 1255569621 M * riotfish Bertl: heh heh i'm already back with questions 1255569622 M * riotfish :) 1255569667 M * riotfish Bertl: my "make menuconfig" doesn't look anything like the one at http://linux-vserver.org/Installation_on_Linux_2.6 1255569726 M * riotfish is there a reason for that ? 1255569741 M * riotfish ahh 1255569745 M * riotfish now i see 1255569759 M * riotfish there are some settings that are version specific :) 1255569848 M * Bertl if you have a working kernel config, it's probably the simplest to copy it into your new kernel tree and run 'make oldconfig' 1255569874 M * Bertl you get then asked about new options and can read the 'help' easily and answer them appropriately 1255569983 M * riotfish Bertl: but if some of the questions are like totally foreign to me, it "should be safe" to just use the default right? 1255570070 M * riotfish how many options should be new when going from 2.6.28-15 to 2.6.31.4 ? 1255570081 M * mugwump 3.2 zillion 1255570092 M * riotfish k... 'cause that's what i'm getting 1255570096 M * riotfish :P 1255570121 M * mugwump I've given up on 'make oldconfig', I just use 'make menuconfig' and exit 1255570125 M * mugwump the defaults are usually fine 1255570137 M * mugwump at least, on the balance of time spent vs reward 1255570147 M * riotfish mugwump: heh heh yeah i've just done the same :) 1255570220 M * riotfish anyone know how long a compile takes on a intel core i7-920 Quadcore ? 1255570229 M * riotfish with 8 gigs of ram 1255570253 M * riotfish like a estimate 1255570839 M * Bertl depends on the kernel config, if you trimmed it down to the essentials, about 2 minutes I'd guess 1255570852 M * mugwump use make -j8 1255570861 M * mugwump or just make -j :) 1255570864 M * riotfish -j8 ? 1255570866 M * Bertl if you do it the distro style and compile in the kitchensink, you might end up with an hour :) 1255570890 M * riotfish Bertl: heh heh then i'm probably gonna end up with an hour :P 1255571200 M * Bertl well, it usually pays off trimming a kernel down for a specific hardware (custom tailoring), this way you reduce the build time (and thus the time on updates) and increase stability (things you don't have compiled in, can't cause panics :) and performance (stuff not running doesn't consume cpu) 1255572457 M * Bertl off to bed now ... have a good one everyone! 1255572464 N * Bertl Bertl_zZ 1255572684 Q * geb Ping timeout: 480 seconds 1255573795 M * riotfish hmm 1255573806 M * riotfish any idea where i can find mkinitrd for ubuntu jaunty 64bit ? 1255573905 M * daniel_hozac try mkinitramfs 1255573926 M * riotfish daniel_hozac: does that suffice as a initrd.img ? 1255573948 M * daniel_hozac yes. 1255573954 M * riotfish cool thnx 1255573954 M * riotfish :) 1255574242 M * riotfish yes! 1255574257 M * riotfish it works (well it boots into the new kernel) 1255574505 M * riotfish isn't it a bad sign if the ./configure of util-vserver reports : checking for vserver... no 1255574946 M * riotfish argh 1255574976 M * riotfish i'm having a problem with dietlibc 1255574989 M * riotfish there is no amd64 version of it for ubuntu! 1255575216 M * riotfish what do i do when i can't have dietlibc? 1255575473 M * riotfish well now i'll just install from source 1255575538 M * riotfish which off course ends in a segmentation fault! 1255575540 M * riotfish grr... 1255575616 J * saulus_ ~saulus@c192178.adsl.hansenet.de 1255575748 Q * hparker Quit: Read error: 104 (Peer reset by connection) 1255576024 Q * SauLus Ping timeout: 480 seconds 1255576028 N * saulus_ SauLus 1255581711 Q * awk 1255581727 J * awk ~phillip@gw1.security.web.za 1255581829 M * awk _Shiva_: well my verbosity logging is perfect. 1255583808 J * Shinsaku ~Shinsaku@chello089076140236.chello.pl 1255584604 Q * balbir_ Ping timeout: 480 seconds 1255585774 J * sharkjaw ~gab@90.149.121.45 1255586945 Q * scientes Remote host closed the connection 1255587769 Q * aj__ Ping timeout: 480 seconds 1255588900 J * aj__ ~aj@80.85.196.112 1255589339 J * scientes ~scientes@174-21-201-131.tukw.qwest.net 1255590838 Q * harry_ Ping timeout: 480 seconds 1255591148 J * harry ~harry@d51A461B4.access.telenet.be 1255591339 Q * awk 1255591354 J * awk ~phillip@gw1.security.web.za 1255591794 J * geb ~geb@earth.gebura.eu.org 1255591851 J * thierryp ~thierry@zanzibar.inria.fr 1255592711 Q * scientes Ping timeout: 480 seconds 1255592749 J * todo HydraIRC@89.72.37.91 1255593299 Q * todo Read error: Connection reset by peer 1255593385 J * harobed ~sklein@pda57-1-82-231-115-1.fbx.proxad.net 1255594112 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d 1255594302 J * mrfree ~mrfree@host1-89-static.40-88-b.business.telecomitalia.it 1255594831 J * balbir_ ~balbir@122.248.163.1 1255595023 Q * awk Ping timeout: 480 seconds 1255595037 J * awk ~phillip@gw1.security.web.za 1255595802 J * leechfest ~chatzilla@216.57.76.186 1255596003 P * leechfest 1255596109 N * Bertl_zZ Bertl 1255596113 M * Bertl morning folks! 1255596172 J * leechfest ~chatzilla@sheba-ipv6.crushedhat.com 1255596183 P * leechfest 1255596302 M * awk morning 1255596423 J * crushedhat ~chatzilla@sheba-ipv6.crushedhat.com 1255596928 Q * balbir_ Ping timeout: 480 seconds 1255596947 J * kir ~kir@swsoft-msk-nat.sw.ru 1255596986 Q * sharkjaw Remote host closed the connection 1255597437 M * crushedhat I have a vserver IPv6 question, probably for bonbons 1255597469 M * Bertl okay? 1255597553 M * crushedhat I'm looking at this log: http://irc.13thfloor.at/LOG/2009-08/LOG_2009-08-21.txt 1255597632 M * crushedhat I'm experiencing the bind problem, and am wondering whether this is believed to be fixed in current code 1255597718 J * balbir_ ~balbir@122.248.163.1 1255597747 M * crushedhat I have patched a Fedora 2.6.30.8-64.fc11.x86_64 kernel with patch-2.6.30.6-vs2.3.0.36.14-pre8.diff (slightly modified for 2.6.30.8) 1255597786 M * crushedhat I have simply booted this kernel, and have not set up any vhosts yet 1255597996 M * crushedhat If there is listener on 0.0.0.0:X, an attempt to listen on [::]:X with IPV6_V6ONLY set will fail with EADDRINUSE 1255598029 M * crushedhat Of course, I would expect this to fail if IPV6_V6ONLY was not set on the socket... 1255598088 M * crushedhat A stock kernel works correctly 1255598148 M * crushedhat Unfortunately, nc doesn't seem to have an option to set IPV6_V6ONLY. I noticed the problem when sshd failed to bind to v6 after binding to v4 1255598155 J * mmouse ~mmouse@office.haefft.de 1255598209 M * Bertl is a known issue for recent kernels, will be fixed as soon as bonbons or I get to it ... 1255598235 M * Bertl for now, either go with the stable branch or avoid binding both v4 and v6 1255598477 M * crushedhat I wouldn't mind trying to fix it myself. Do you have any clues about the problem beyond what is discussed in the 08-21 chat (maybe something related to ipv6_rcv_saddr_equal())? 1255598518 M * crushedhat Do you know the latest kernel that doesn't have this issue? 2.6.29? 1255598587 M * Bertl sec, it was a mainline change which caused this .. 1255598718 M * Bertl 2.6.27.x should be fine, 2.6.29 seems to be the last one (date of af_inet6.c.rej) 1255598827 M * Bertl the problem is, the 'checks' moved to ipv6_rcv_saddr_equal() as you already know, and bonbons adapted the code to the changes, but obviously missed a few cases, now my plan is to make the Linux-VServer specific checks very similar to the mainline checks, which should (in theory) work, but needs testing ... 1255598865 M * Bertl for this purpose, I started to build a tool and plan to write a testscript (doing all the possible v4/v6 host/guest scenarios) 1255598909 M * Bertl but so far, I didn't find the time to complete the tool or write the script, thus the code remains unchanged, as it only affects a very small group of users ... 1255598968 M * Bertl so, if you really want to help, either fix it by adding the necessary checks, or alternatively, help with the test script, so that I can do the proper changes and test the results :) 1255598986 M * awk Bertl: hmm, now that I don't have debootstrap, how would I install say 'debian sid' on a system.... vserver sandbox build -m template --context 42 --hostname sandbox --interface eth0:172.17.0.248/24 -- -d debian 1255599002 M * awk that doesn't seem to work... I would like to point my mirror to eg: http://debian.mirror.ac.za/ 1255599022 M * Bertl why not use the debootstrap build method? 1255599036 M * awk Bertl: installing on CentOS host 1255599040 M * Bertl so? 1255599053 M * awk debootstrap is not supplied with centos? 1255599057 M * Bertl so? 1255599065 M * crushedhat I would say this affects anyone using IPv6, since sshd now acts differently even on the host instance. Maybe that's a very small group (sigh...) 1255599068 M * awk heh... am I missing the point? 1255599088 M * awk convert the .deb package to rpm and install like that? 1255599090 M * Bertl debootstrap will be downloaded automatically by util-vserver when needed 1255599103 M * Bertl just use the usual build line 1255599107 M * awk ok, let me try the method and show you why I asked as it never worked.. 1255599113 M * awk thanks 1255599126 M * Bertl you got recent util-vserver, yes? 1255599130 M * crushedhat I will see if I go insane trying to deduce all the cases in the new kernel code 1255599141 M * Bertl crushedhat: yeah, still the number of ipv6 folks is really small 1255599198 M * awk util-vserver-0.30.216-1.pre2793.el5.centos 1255599214 M * Bertl sounds good ... 1255599234 M * _Shiva_ Bertl: but we that already are v6 enabled WILL GROW! ;-) ("resistance is futile") ;-) 1255599253 M * Bertl awk: although pre2848 would be better ... 1255599453 M * awk ERROR: Could not download the debootstrap package from http://ftp.debian.org/debian/pool/main/d/debootstrap/debootstrap_1.0.10_all.deb 1255599461 M * awk thats why it never worked :P 1255599517 M * crushedhat Bertl: Thanks for your help. I mostly wanted to confirm that there wasn't a fix somewhere that I missed before I spent time trying to fix this. Do you think it's likely that bonbons might have some additional clues to get me started? When is he usually around? 1255599517 M * Bertl either get a newer release, which contains the new pathes, or do as suggested, i.e. addjust it 1255599524 M * Bertl (that was for awk :) 1255599528 M * mmouse that package is called debootstrap_1.0.10lenny1_all.deb 1255599539 M * awk Bertl: just editing it, thanks... 1255599588 M * Bertl crushedhat: maybe, but he also said he would take a look at it (to arekm IIRC), so he might be already working on something .. read up on the IRC logs, they contain all we know so far 1255599593 J * gnuk ~F404ror@pla93-3-82-240-11-251.fbx.proxad.net 1255599596 M * awk Bertl works :D 1255599654 M * bonbons crushedhat: no specific additional suggestions, best is to check which cases mainline did apply and the add per-context filters on top of it (or determine what cases I factorized incorrectly) 1255599827 Q * mmouse 1255599857 M * awk sheesh, remeber awhile back was told to stop using debootstrap 1255599866 J * sharkjaw ~gab@90.149.121.45 1255599898 M * Bertl awk: who told you that? 1255599914 M * crushedhat Good to see you bonbons! I’ll see what I can do, and send you a patch of course if I fix it. 1255599943 M * Bertl paste it somewhere and drop an url here, if you got something tested 1255599971 M * awk hmmm, few years back was running into problems with debootstrap and somebody said, don't use that and rather use some or other way of building 1255599999 M * Bertl for debian/ubuntu guests, it is the preferred method 1255600008 M * Bertl so ignore what somebody said back then 1255600011 M * daniel_hozac always has been 1255600019 M * awk sure 1255600024 M * Bertl off for now ... some work is waiting ... 1255600029 N * Bertl Bertl_oO 1255600833 M * awk Hi, where I can I read the channel history? 1255600839 M * awk need to see something I asked yesterday? 1255600892 M * BWare check www.linux-vserver.org 1255600908 M * BWare there should be a link to the irc logs somewhere 1255600945 M * crushedhat The IRC logs are at http://irc.13thfloor.at/LOG/ 1255600987 P * crushedhat 1255600996 J * crushedhat ~chatzilla@sheba-ipv6.crushedhat.com 1255601014 Q * crushedhat Quit: ChatZilla 0.9.84 [Firefox 3.0.14/2009082707] 1255601033 J * crushedhat ~tim@sheba-ipv6.crushedhat.com 1255601118 M * awk Thanks! 1255601570 J * hparker ~hparker@2001:470:1f0f:32c:290:96ff:fe50:40fa 1255601579 J * ghislainocfs21 ~Ghislain@adsl2.aqueos.com 1255601788 Q * balbir_ Ping timeout: 480 seconds 1255601934 Q * ghislainocfs2 Ping timeout: 480 seconds 1255602573 Q * ghislainocfs21 Ping timeout: 480 seconds 1255602893 M * awk shoe, ok been doing allot of reading and can't find if I have a centos host, now I use the debootstrap install, yet I have 0 packages... I need to install something like apt so I can at least start... yet, how do I do this from the guest, seeing its not rpm based so I can vyum install .... 1255603182 J * JonB ~NoSuchUse@130.227.63.19 1255603219 M * JonB daniel_hozac: after upgrading to latest tarball, i can not start my guests. Is vprocunhide not used any more? 1255603281 M * riotfish why doesn't dietlibc exist for 64bit ? 1255603296 M * riotfish (needed for util-vserver) 1255603377 M * JonB daniel_hozac: never mind 1255603510 M * riotfish Bertl_oO: do you have any suggestions to how to get util-vserver installed when I can't get dietlibc installed 'cause of 64bit arch? 1255603607 M * hparker riotfish: I run vserver 64bit 1255603644 Q * JonB Quit: Leaving 1255603689 M * riotfish hparker: how did you install util-vserver? (i'm using ubuntu 9.04 64bit) 1255603708 M * hparker I emerged it (Running Gentoo) 1255603720 M * hparker No clue on Ubuntu, never seen it 1255603730 M * riotfish ahh, that doesn't help me very much :) 1255603743 Q * fback Read error: Connection reset by peer 1255603758 J * fback fback@red.fback.net 1255603765 M * hparker Sure it does, it lets you know it's at least possible.. Now you just need to figure out how ;) 1255603767 M * riotfish apparently there isn't any 64bit version of dietlibc for ubuntu jaunty nor debian lenny... 1255603775 M * riotfish hparker: heh heh yeah that's right 1255603796 M * hparker Worst case, grab the tarball and compile it 1255603810 M * riotfish i'm trying to compile dietlibc, but i get some wierd message about some: make: *** [bin-x86_64/elftrunc] Segmentation fault 1255603833 M * hparker eww 1255603837 M * hparker No clue, sorry 1255603842 M * riotfish hparker: fair deal :) 1255603935 M * riotfish just to make it clear this is what it fails on: 1255603938 M * riotfish bin-x86_64/diet gcc -D__dietlibc__ -Os -fstrict-aliasing -momit-leaf-frame-pointer -mfancy-math-387 -W -Wall -Wextra -Wchar-subscripts -Wmissing-prototypes -Wmissing-declarations -Wno-switch -Wno-unused -Wredundant-decls -o bin-x86_64/elftrunc contrib/elftrunc.c 1255604059 M * riotfish hmm... looks like i fixed it 1255604063 M * riotfish nope 1255604288 Q * manana Ping timeout: 480 seconds 1255604309 P * halfd 1255604459 M * riotfish oh well, there is a .deb package 1255604470 M * riotfish for ubuntu jaunty 1255604477 M * riotfish util-vserver that is 1255604478 M * riotfish :) 1255604567 M * hparker There ya go 1255604628 M * hparker What version? 1255604822 J * manana mayday_270@84.17.25.144 1255604982 M * riotfish hparker: 0.30.216~r2772 1255605005 M * hparker k 1255605011 M * riotfish but now i have another problem... i don't have any vprocunhide or vservers-default in my /etc/init.d/ 1255605019 M * hparker :P 1255605037 M * riotfish and where the f*ck do i get those from... 1255605053 M * hparker I would think they'd be in the package 1255605072 M * riotfish hmm... yeah i would to 1255605105 M * riotfish and vprocunhide is, but it gets placed at /usr/lib/util-vserver/vprocunhide and not /etc/init.d/ 1255605271 M * hparker Hopefully someone with Ubuntu experience will come along and help you 1255605272 M * riotfish oh well 1255605287 M * riotfish the testscript from http://linux-vserver.org/Installation_on_Linux_2.6 works all the way 1255607903 J * doener ~doener@i59F5472B.versanet.de 1255608008 Q * doener_ Ping timeout: 480 seconds 1255608632 J * Medivh ck@dolphin.serverbox.de 1255610107 Q * harobed Quit: Ex-Chat 1255611959 J * BenG ~bengreen@94-169-110-10.cable.ubr22.aztw.blueyonder.co.uk 1255611970 M * BenG hey all 1255611989 M * BenG looks like there will probably be a vserver patch in Debian Squeeze 1255612115 Q * Hunger Remote host closed the connection 1255612379 M * BenG hurrah! 1255612420 M * hparker Going to squeeze it in? :P 1255612839 M * BenG :) 1255612897 Q * BenG Quit: I Leave 1255613155 Q * jrdnyquist Quit: by 1255614396 M * awk Is there a way I can change it so it doesn't show /dev/hdv ? 1255614399 M * awk but something else? 1255614408 M * awk seeing some patches coming out checking that if / is vserver 1255614428 M * awk change device name? 1255615134 Q * Mr_Smoke Read error: Connection reset by peer 1255615153 J * Mr_Smoke ~smokey@layla.lecoyote.org 1255615201 Q * sharkjaw Remote host closed the connection 1255615349 M * awk ok never mind worked it out 1255615526 M * awk hrm, each time I edit /vservers/name/etc/mtab and restart it goes back to hdv1 1255615531 M * awk I need to modify this... 1255615577 M * daniel_hozac /etc/vservers//apps/init/mtab 1255615579 Q * SauLus Ping timeout: 480 seconds 1255615683 M * awk yay :D 1255615687 M * awk thanks.. 1255616567 J * Hunger ~Hunger@Hunger.hu 1255616923 Q * geb Ping timeout: 480 seconds 1255617203 Q * aj__ Ping timeout: 480 seconds 1255617512 Q * thierryp Quit: ciao folks 1255617617 J * geb ~geb@173.4.82-79.rev.gaoland.net 1255618128 J * dholbach ~daniel@i59F72D8A.versanet.de 1255618132 M * dholbach hi 1255618230 M * dholbach does anybody of you know how to best deal with vserver guests that run ubuntu karmic (9.10)? 1255618242 M * dholbach is there a best-practise for that already? :) 1255618969 Q * mrfree Ping timeout: 480 seconds 1255619090 J * aj__ ~aj@80.85.196.112 1255620347 J * urbee ~dssd@internet-213-229-247-101.narocnik.mobitel.si 1255620349 M * urbee hoi hoi =) 1255620367 M * urbee i'm trying to set iptables on a guest server, this is what i get when i do 'iptables -L' 1255620372 M * urbee FATAL: Could not load /lib/modules/2.6.29-vs2.3.0.36.14-gentoo/modules.dep: No such file or directory 1255620380 M * urbee iptables v1.4.2: can't initialize iptables table `filter': Permission denied (you must be root) 1255620386 M * urbee Perhaps iptables or your kernel needs to be upgraded. 1255620403 M * hparker urbee: From within the guest? 1255620416 M * urbee yes 1255620507 M * hparker No 1255620512 M * urbee no what :P 1255620514 M * hparker Iptables on the host 1255620529 M * urbee iptables on the host what? 1255620530 M * urbee :) 1255620540 M * urbee i'm tryin this from within guest vserver 1255620545 M * urbee one of my clients needs to close some ports 1255620593 M * hparker You do it on the host, not within a guest 1255620602 M * urbee so noone on guest can use iptables? 1255620609 M * hparker right 1255620670 M * urbee huh 1255620683 M * urbee any special command to do that 1255620700 M * urbee or do i just use the prefered along the command 1255620738 M * urbee prefered ip* 1255620824 M * dholbach I realise I was a bit unclear with my question a few minutes ago... I was wondering if there's a best-practise way to deal with upstart in ubuntu karmic guests? (https://bugs.launchpad.net/ubuntu/+source/upstart/+bug/430224 seems related to the problems I'm seeing) 1255622084 Q * urbee Ping timeout: 480 seconds 1255622814 P * ncopa Ex-Chat 1255623061 J * nkukard_ ~nkukard@196.212.73.74 1255623252 Q * geb Ping timeout: 480 seconds 1255623263 J * geb ~geb@earth.gebura.eu.org 1255623426 Q * nkukard Ping timeout: 480 seconds 1255623578 Q * kir Quit: Leaving. 1255623996 N * Bertl_oO Bertl 1255624030 M * Bertl dholbach: I'd say, you want to go for the 'plain' init style and configure upstart properly? 1255624096 M * dholbach the problem is that there's the first few daemons that ship only upstart jobs 1255624104 M * dholbach and not "old style" init scripts any more 1255624125 M * Bertl a quite bad idea, IMHO, as it forces you to run an init process 1255624127 M * dholbach is it generally possible for a guest to have its own PID 1 and /dev/console? 1255624132 M * Bertl yep 1255624163 M * dholbach how would I go about doing that? 1255624167 M * Bertl default is 'sysv' init style, where no init is started inside the guest, but 'plain' style will start whatever init process you specify 1255624177 M * dholbach it'd be nice if we could at least document something, so karmic is not going to screw all vserver users 1255624189 M * dholbach ah! 1255624208 J * vorlon vorlon@minbar.dodds.net 1255624215 M * dholbach hi vorlon 1255624217 M * Bertl go ahead, test it out, write down what you needed to do/cleanup/etc and put it on the wiki 1255624227 A * vorlon waves 1255624227 M * dholbach Bertl: just said that it should generally be possible: 1255624229 M * dholbach default is 'sysv' init style, where no init is started inside the guest, but 'plain' style will start whatever init process you specify 1255624251 M * vorlon right, that sounds like what you need 1255624276 M * dholbach I'll talk to the guy who can do stuff "out side the guest" 1255624277 J * jrdnyquist ~jrdnyquis@slayer.caro.net 1255624293 M * dholbach vorlon: you said something about /dev/console too 1255624323 M * vorlon yeah, if upstart can't open /dev/console, it's probably going to be unhappy 1255624335 M * vorlon (this is just a guess btw, I haven't looked at the code) 1255624369 M * dholbach Bertl: Gonzo says hi and asks where to set the "plain" style? is there some kind of documentation? 1255624406 M * Bertl http://www.nongnu.org/util-vserver/doc/conf/configuration.html 1255624416 M * dholbach Bertl: do you know anything about guests being able to open /dev/console or something? 1255624417 M * Bertl (as usual, the great flower page) 1255624433 M * Bertl well, if there is an /dev/console, a guest can open it :) 1255624474 M * dholbach sorry for being so dense, but as you probably find out, I have no clue and am merely trying to ask the right things :) 1255624480 M * dholbach Bertl: thanks a lot for all that already 1255624493 M * dholbach I hope we can get something into the Ubuntu release notes, so people know what's happening 1255624508 M * dholbach although a lot will probably have to contact their ISPs to change their vserver configuration :/ 1255624551 M * Bertl I don't think ubuntu is that common for Linux-VServer guests 1255624578 M * Bertl but yeah, it's good to document is somewhere 1255624590 M * Bertl (preferably on the wiki, that is) 1255624604 M * dholbach right 1255624605 M * Bertl and greetings back to Gonzo 1255624611 M * dholbach make it upstart-using-distro guests then :) 1255624641 M * fback good evening folks :) 1255624671 Q * aj__ Ping timeout: 480 seconds 1255624798 J * gonzo ~sh@80.69.41.3 1255624806 M * dholbach gonzo! 1255624810 A * dholbach hugs gonzo 1255624816 M * gonzo hi daniel, hi bertl, hi rest 1255624822 M * Bertl hi gonzo! 1255624870 Q * Shinsaku 1255624919 M * gonzo I echoed "plain" into /etc/vservers//apps/init/style 1255624924 M * gonzo and started it 1255624947 M * gonzo actually it just started 'init' then, nothing else 1255624988 M * gonzo ah... I see, it's not switching to a runlevel 1255625015 M * dholbach vorlon: ^ do you know if anything else should happen then 1255625037 M * gonzo dholbach: which runlevel would you prefer? 2? 1255625075 M * dholbach I don't know 1255625090 M * gonzo I'll try 2 and 6 for stopping 1255625093 M * dholbach gonzo: what do you need to do to make that happen? just so we can document it afterwards 1255625197 M * gonzo I managed to start your vserver with your custom init and then I have to enter it, to run "init 2" 1255625219 M * dholbach :( 1255625221 M * gonzo putting "2" into apps/init/runlevel did not help to automize that prcess 1255625290 M * Bertl with the plain init style, your init process is started 1255625306 M * Bertl there is no runlevel handling from outside anymore, just signalling 1255625404 M * Bertl try to specify a bootflag if upstart needs that 1255625416 M * gonzo two problems there: 1) it's not switching to the desired runlevel by itself 2) the %!§&"!§ upstart init doesn't work, too. It still does not start syslog, cron, etc. Just a few services using sysv init wrappers 1255625440 M * gonzo bootflag? by using "plain "? 1255625502 M * Bertl nah, at least I don't think util-vserver supports that 1255625520 M * gonzo ok :) 1255625553 M * Bertl is there any useful upstart man page online? 1255625558 J * urbee ~dssd@internet-213-229-238-74.narocnik.mobitel.si 1255625561 M * urbee hi 1255625565 M * Bertl hey 1255625586 M * urbee i'm still having problems setting firewall on a guest vserver 1255625609 M * Bertl not unusual, you set it on the host ... :) 1255625645 M * urbee are there any iptables rules? 1255625664 M * Bertl you tell me, I don't know your setup :) 1255625686 M * Bertl but maybe let's take a step back, and describe it, and what you want to accomplish? 1255625687 M * gonzo the upstart manpage is very basic. I could pastebin it, if you like 1255625693 M * Bertl please do so 1255625715 M * urbee I'd like to close all ports except 2 of them on a guest vserver 1255625749 M * Bertl okay, and the guest has a public IP? 1255625754 M * urbee yes 1255625761 M * Bertl different from the host IP? 1255625775 M * urbee well..yes 1255625790 M * urbee same class, different ip 1255625796 M * gonzo @Bertl: http://pastebin.com/m658cd2f3 1255625814 M * Bertl urbee: good, and you do not want to affect the host IP? or close all ports on the host too? 1255625823 M * urbee no, i dont want to affect host ip 1255625824 M * urbee just the gues 1255625825 M * urbee t 1255625853 M * Bertl okay, what about host-guest traffic? 1255625871 M * urbee that doesnt matter 1255625882 M * urbee outside traffic is important 1255625905 M * Bertl well, then add an iptables rule like this: 1255625956 J * aj__ ~aj@c193034.adsl.hansenet.de 1255625983 M * Bertl iptables -A INPUT -d -j ACCEPT 1255625998 M * Bertl for all inbound ports, and a blocking rule like: 1255626012 M * Bertl iptables -A INPUT -d -j DROP/REJECT 1255626029 M * Bertl at the end, same for output if you want to block outgoing traffic too 1255626051 A * gonzo greets aj__. 1255626057 M * urbee life saviour :) 1255626128 M * Bertl gonzo: unfortunately that man page tells nothing (at least nothing useful) 1255626139 M * gonzo Just like I said :) 1255626146 M * Bertl gonzo: the questions which need to be answered are these: 1255626147 J * Pazzo ~ugelt@reserved-225136.rol.raiffeisen.net 1255626164 M * Bertl a) how is upstart started from initramfs on a normal system? 1255626189 M * Bertl b) how is upstart signalled to shutdown/reboot 1255626288 M * Bertl c) how to specify a console, but that is maculation 1255626351 M * urbee box5 ~ # iptables -A INPUT -d 91.185.208.2 TCP 80 -j ACCEPT 1255626351 M * urbee Bad argument `TCP' 1255626394 M * Bertl man iptables :) 1255626470 M * Bertl iptables -A INPUT -d 91.185.208.2 -p tcp --dport 80 -j ACCEPT 1255626495 M * Bertl (you might also add the interface with -i 1255626512 M * urbee i was just manning that! :) 1255626551 M * Bertl good, man-on :) 1255626552 M * urbee iptables -A INPUT -d 91.185.208.2 -p tcp --dport 80 -i eth0 -j ACCEPT 1255626560 M * urbee seems to work :p 1255626720 Q * Pazzo Quit: Bye! 1255627050 M * _Shiva_ Bertl: reading about iptables-rules.. might the upcoming nftables/nflib be interfacable from a guest? 1255627300 M * _Shiva_ Bertl: like: the host's admin defines some returning queue hook for the guest's nid and the guest's root is able to add some rules w/i his namespace? 1255627401 M * Bertl maybe, if that becomes interesting for users and/or development gets sponsored (or somebody digs into it) I could imagine using that for Linux-VServer 1255627445 M * _Shiva_ Bertl: http://marc.info/?l=linux-kernel&m=122363656415972&w=2 1255627467 M * _Shiva_ :-) 1255627502 M * _Shiva_ . o 0 ( LXC - yet again ;-) ) 1255627718 M * Bertl I don't get the relation, this posting is quite old .. care to explain? 1255627806 M * _Shiva_ apart from the date - i thought LXC was mainly trated as the "evil sister of linux-vserver".. at least in this channel ;-) 1255627836 Q * urbee Ping timeout: 480 seconds 1255627851 M * Bertl nah, not really, LXC is the future ... 1255627903 M * _Shiva_ but from googling for more recent information on the nftables topic - this seems to be the first and latest (official) sign of life 1255627914 M * Bertl (it's just not usable yet, despite the hype :) 1255627949 M * Bertl http://lwn.net/Articles/324251/ 1255627997 M * _Shiva_ "With the kernel 2.6.29, lxc is fully functional." (<- that's from http://lxc.sourceforge.net/lxc.html dated Fri Jul 24 16:23:02 CEST 2009) 1255628001 M * _Shiva_ *g* 1255628019 M * Bertl as I said, hype ... I hope that is not the full potential of LXC :) 1255628212 M * fback Bertl: you're using kvm right? 1255628244 M * fback Bertl: is it possible to install winXP as a guest system at this stage? 1255628325 M * _Shiva_ fback: i'm running even Win7 64bit on KVM.. ;) 1255628449 M * Bertl fback: no idea, why would I install winXP :) but yeah, I'd say it should work pretty fine :) 1255628487 M * Bertl (well, as fine as microsoft products run :) 1255628534 M * Bertl gonzo: maybe contact the upstart developer(s) or check the source code? 1255628591 M * gonzo I am checking... the documentation, even in the wiki / FAQ seems to be.... more into "howto change your init scripts in order to use upstart" 1255628615 M * gonzo the developer was contacted by dholbach, no response yet, AFAIK 1255628710 M * fback Bertl: I have windows-olny laser printer from canon 1255628763 M * fback Bertl: there are drivers only for 32-bit winXP 1255628828 M * Bertl sell it on ebay :) 1255628862 M * Bertl not worth the trouble ... 1255628931 M * fback Bertl: it's not mine, but my company 1255628949 M * gonzo fback: Sell your company on ebay - not worth the trouble :-p 1255629088 Q * riotfish Quit: Leaving 1255629100 M * fback strange thing, I had to power-cycle the computer to enable cpu virtualization extensions, simple reboot wasn't enough 8) 1255629205 M * _Shiva_ fback: some Intel docs even describe to power off (and disconnect the power cord) for at least a min when changing VT-? settings... 1255629239 M * _Shiva_ fback: have read that for enabling VT-d on supporting north bridges 1255629321 M * _Shiva_ . o 0 ( and if virtualization is such a hype - why is that thingy always diabled by default? ) 1255629324 M * Bertl as long as you do not have to bubble wrap the cpu :) 1255629528 M * fback _Shiva_: that's what I had to do 1255630657 J * BenG ~bengreen@94-169-110-10.cable.ubr22.aztw.blueyonder.co.uk 1255631051 J * imcsk8 ~ichavero@148.229.1.11 1255632005 M * _Shiva_ fback: and that's what I meant - it's _always_ disabled by default.. on any machine that i've seen so far 1255632367 Q * BenG Quit: I Leave 1255632517 Q * gnuk Quit: NoFeature 1255632762 Q * geb Ping timeout: 480 seconds 1255633012 Q * dholbach Quit: Ex-Chat 1255633164 J * Pazzo ~ugelt@sadsl-246059.rol.raiffeisen.net 1255633570 J * geb ~geb@173.4.82-79.rev.gaoland.net 1255633990 Q * bonbons Remote host closed the connection 1255634293 J * derjohn_foo ~aj@c207029.adsl.hansenet.de 1255634709 Q * aj__ Ping timeout: 480 seconds 1255635665 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d 1255636368 Q * Bertl Ping timeout: 480 seconds 1255636524 J * Bertl herbert@IRC.13thfloor.at 1255637742 Q * geb Ping timeout: 480 seconds 1255638540 J * geb ~geb@earth.gebura.eu.org 1255638619 J * Alteisen alteisen@shell.chaostreff-dortmund.de 1255638634 M * Alteisen hi everyone 1255638788 M * Alteisen i have a question about external usb harddisks and vserver - i want to share an external 1 TB usb drive from within a vserver guest via samba. which filesystem is recommended? ntfs or ext3? are there any howtos doing this? 1255638806 M * daniel_hozac vserver doesn't really influence the filesystem choice. 1255638940 M * Alteisen so it is ok to use ntfs via ntfs-3g on the external drive? might be helpful, when attaching the drive directly to a windows host. 1255638968 M * daniel_hozac possible, sure. 1255638981 M * daniel_hozac obviously ext3 works better on Linux hosts. 1255638987 M * daniel_hozac but, whatever floats your boat. 1255639023 M * Alteisen what will happen to the vserver guest, when i start it with the external drive still switched off? 1255639052 M * Bertl one thing regarding the filesystem, ntfs does not support the Linux-VServer specific attributes 1255639080 M * Bertl so, you won't be able to use unification or tagging ... but that shouldn't be a big deal 1255639086 M * Alteisen does it matter, if i use that usb drive exclusivly for one vserver guest? 1255639124 M * Bertl don't think so .. but, wasn't ntfs write support experimental? 1255639155 M * Alteisen ntfs-3g was mentioned "stable" in the internet 1255639170 M * Bertl and with experimental, I actually mean broken in some weird way :) 1255639173 M * Alteisen also for write 1255639190 M * Bertl anyway, do some tests first ... before you lose data 1255639271 M * Alteisen ext3 sounds less dangerous... 1255639329 M * Alteisen but moving the gigs of data from that ntfs drive to somewhere else, converting to ext3 and moving back the files, means lots of work and time.. 1255639355 M * daniel_hozac ntfs-3g is a completely different way of doing it. 1255639365 M * daniel_hozac it has "solid" write support. 1255639393 M * Bertl ah, okay, forget what I said then :) 1255639754 M * Alteisen well, i'll give ntfs-3g a try 1255639825 M * Bertl please let us know how it goes ... 1255639890 M * Alteisen on debian, it looks quite simple. apt-get install ntfs-3g, then mount -t ntfs-3g /dev/sdc1 /mnt 1255639915 M * Alteisen in /mnt, you can create directorys, copy files... 1255640165 M * Bertl hmm ... probably needs a kernel module, no? 1255640173 M * daniel_hozac no. 1255640176 M * daniel_hozac FUSE. 1255640183 M * Bertl ah, cool then 1255640272 M * Alteisen fuse is a kernel module 1255640334 M * Bertl yeah, but mainline 1255640475 M * Alteisen why do you separate /etc/vservers from the directory, where the vserver files are? wouldn't it be easier, to have an "etc" or ".config" within the vserver directory? 1255640502 M * Bertl self modifying guests, not a good idea :) 1255640526 M * daniel_hozac the configuration directory cannot live below the barrier. 1255640549 M * Bertl well, the barrier is 'almost' gone :) 1255640564 M * Alteisen but the vserver can not breakout from it's "chroot"?! 1255640573 M * daniel_hozac because of the barrier. 1255640577 M * daniel_hozac (in previous versions) 1255640588 M * daniel_hozac chroot is incredibly easy to break out of. 1255640658 M * Alteisen my question is only about backup of the vservers, which means i have to backup my /vservers directory and the /etc/vservers directory 1255640674 M * Bertl do you change the config that often? 1255640684 M * Alteisen only one place for both would have been nice 1255640696 M * Alteisen Bertl: hm, good question 1255640698 M * Bertl for backups, maybe :) 1255640701 M * daniel_hozac well, two paths instead of one is not really difficult. 1255640823 M * Alteisen ok, if there are (or were) technical reasons, to separate configs and files that way, that is acceptable 1255640833 M * Alteisen i just was questioning that 1255640857 M * daniel_hozac there are reasons every service you have on your systems keep their configs separate from their data. 1255640859 M * Bertl I'd sugguest to backup /etc/vservers once a day or maybe only when you change the config 1255640868 M * Alteisen if there is any reason, also historical, that is acceptable for me 1255640889 M * daniel_hozac you wouldn't want apache to store it's configuration in the web root, would you? 1255640947 M * Alteisen i do my backups on a "per vserver" base, including the /etc/vserver/name and /vservers/name directory in one step. vserver stop, rsync, vserver start. 1255641069 M * Alteisen what is wrong with /vservers/ and /vservers/etc? can a vserver access files above /vservers/? 1255641103 M * Bertl ah, now I understand where you are heading to 1255641107 M * daniel_hozac what would the difference be? 1255641115 M * Bertl but that won't help you with the backups, would it? 1255641151 M * Alteisen i might backup /vservers instead of /vservers and /etc/vserver 1255641164 M * Bertl on a per guest basis? 1255641202 M * Alteisen no, that would be on the "global" basis 1255641224 M * Alteisen all vservers - and the config - in one step 1255641242 M * Bertl well, then you can do it right now in only two steps :) 1255641306 M * Alteisen okay :-) 1255641395 M * Alteisen as i am still learning linux and vserver, while administrating windows based networks in my job, i am maybe blinded and confused 1255641411 M * Alteisen but that is no excuse 1255641419 M * Bertl np, we'll open your eyes and teach you the ways :) 1255641498 M * Alteisen a long time ago, i found this page: http://lena.franken.de/linux/debian_and_vserver/vserver_fastrestore.html 1255641540 M * Alteisen i am still trying to get it working that way, but even easier and in a way, i can understand it *g* 1255641627 M * Alteisen and i don't want to use "ghost" for backing up my host system 1255641640 M * Bertl there are alternatives 1255641665 M * Alteisen i have a preseed setup running for my host 1255641670 M * Bertl http://www.mondorescue.org/ 1255641819 M * Alteisen so i just have to choose pxe boot, select my server's name in the menu, wait for max 40 minutes, and then restore /etc/vservers and /vservers and i am done. 1255641864 M * Alteisen all i need, is a computer with at least 256 MB RAM, a x86 processor and pxe boot option *g* 1255642046 M * Alteisen (and my dhcp/tftpboot environment, but that is on a dsl-router without spinning parts) 1255642199 J * blues ~blues@acty1.neoplus.adsl.tpnet.pl 1255642319 Q * blues_ Ping timeout: 480 seconds 1255642350 M * Alteisen time to sleep now 1255642355 M * Alteisen good night 1255642780 M * Bertl have a good one! 1255644770 J * carnage ~carnage@voip-colo-74-86-148-74.link2voip.com 1255646771 Q * FireEgl Read error: Operation timed out 1255647484 J * FireEgl FireEgl@173-16-9-10.client.mchsi.com 1255647910 J * ksn ~kuben@82.128.68.122 1255648822 Q * awk Read error: Connection reset by peer 1255648839 J * awk ~phillip@gw1.security.web.za 1255649110 Q * imcsk8 Quit: Leaving 1255649161 M * vorlon Bertl: upstart is started from initramfs via 'exec init'; should not need any arguments, it looks at the kernel commandline instead (/proc/cmdline) for its options; b) upstart provides its own shutdown/reboot/halt commands which talk to it using the dbus protocol (but I believe the init daemon can also process traditional 'telinit' commands for upgrade compatibility); c) init expects /dev/console to be what it wants/needs, AFAICT 1255650058 Q * geb Ping timeout: 480 seconds 1255650119 J * saulus ~saulus@d047127.adsl.hansenet.de 1255650316 Q * bonbons Quit: Leaving 1255650344 M * Bertl vorlon: okay, so one needs to fake /proc/cmdline to satisfy upstart? 1255650382 M * Bertl the shutdown/reboot/halt is fine, but what about traditional init signalling? 1255650443 M * Bertl what does it do when sent an SIGINT (normally init executes the ctrlaltdel action) 1255650479 Q * saulus Remote host closed the connection 1255650513 M * Bertl regarding console, that should be doable, although I consider it less than optimal that it cannot be configured (as for any init) 1255650641 M * vorlon Bertl: I expect /proc/cmdline needs faked, yes; and SIGINT is handled 1255650867 M * Bertl okay, and without any proper /proc/cmdline, upstart just sits around doing nothing? 1255650888 M * vorlon no, it launches udev and mounts filesystems ;) 1255650934 M * Bertl but it doesn't start any services ... well, interesting design :) 1255650966 M * vorlon (really, by default I would expect it to enter runlevel 2 if there was no /proc/cmdline; this is all defined in the upstart jobs, not hard-coded in init; if it didn't switch to runlevel 2, I suspect it's gotten confused at the filesystem-mounting stage) 1255651018 M * vorlon in a current karmic system, this is /etc/init/rc-sysinit.conf 1255651029 M * vorlon # Default runlevel, this may be overriden on the kernel command-line 1255651029 M * vorlon # or by faking an old /etc/inittab entry 1255651029 M * vorlon env DEFAULT_RUNLEVEL=2 1255651039 M * Bertl hmm, let's hope it is the /dev/console part which confuses it, because the filesystem mounting is not an option 1255651071 M * Bertl and would render upstart useless for any container system 1255651074 M * vorlon are /proc, /sys, etc. pre-mounted? 1255651092 M * Bertl yes, filesystems are mounted from the host 1255651106 M * vorlon and does /proc/mounts look sensible within the container? 1255651109 J * saulus ~saulus@d047127.adsl.hansenet.de 1255651117 M * Bertl yes, it should 1255651134 M * vorlon if so, mountall (upstart's mount helper) should figure out from there that it doesn't need to do anything - and if not, that's a bug 1255651178 M * vorlon editing /etc/init/mountall.conf to call 'mountall --debug' and capturing the output somewhere would probably be useful