1134951451 P * meandtheshell 1134951565 J * ryker ~ryker@c-67-176-240-74.hsd1.in.comcast.net 1134951892 J * Johnnie ~jdlewis@acs-24-154-53-16.zoominternet.net 1134953189 J * meandtheshell ~itsmethem@85-124-8-44.dynamic.xdsl-line.inode.at 1134953295 M * jayeola hey guys, does one have "CONFIG_VSERVER_NGNE" enabled or not? the text in the ncurses help file is slightly confusing 1134953349 J * shedi ~siggi@inferno.lhi.is 1134953394 M * derjohn jayeola, what is the text like? I wonder if "CONFIG_VSERVER_NGNE". If this is the case it should be regardrd as experimental (dont use).. what kernel do you use? VS version? patches? 1134953415 M * derjohn I wonder if "CONFIG_VSERVER_NGNE" id "NGNET" 1134953420 M * derjohn id=0is 1134953430 M * derjohn (sry ... i am tired) 1134953433 M * jayeola distro centos 4.2, Linux Kernel v2.6.14.3-vs2.0.1 Configuration 1134953475 M * jayeola heh - i'm tired and hungrybut this box has got to build. 1134953497 M * derjohn Dr. Hungrybut ;) 1134953503 M * derjohn well 1134953519 M * jayeola "CONFIG_VSERVER_NGNET" , yes you where right... 1134953524 M * derjohn gg says you are talkung about "CONFIG_VSERVER_NGNET" not "CONFIG_VSERVER_NGNE" 1134953524 M * derjohn ah 1134953530 M * derjohn same idea ;) 1134953543 M * jayeola sloppy yanking of text 1134953553 M * jayeola on my part... 1134953577 M * derjohn say the kernel help "Next Generatio networking" ? 1134953689 M * jayeola this help? http://pastebin.com/469611 1134953698 M * derjohn I wonder why NGNET is in stable VS2 .. but you wont need ngnet. http://linux-vserver.org/NGNET-Development 1134953750 M * derjohn ah 1134953751 M * Doener jayeola: well, the text says "Do not disable unless you exactly know what you are doing." ;) 1134953799 M * jayeola heh - yah, so if i make it a *, does that mean that i've disabled it? that's what i need clearing up 1134953800 M * Doener and that means the legacy network stuff which is disabled by the switch --> have CONFIG_VSERVER_NGNET disabled 1134953818 M * Doener (yeah, it _is_ confusing ;) 1134953843 M * jayeola i understand (roughly) the concept, it's just the wording of that text 1134953862 M * derjohn yes, it is confusing. sould read: CONFIG_VSERVER_DISABLE_NGNET then 1134953965 M * jayeola thanks for your time bty.... 1134954212 M * Doener derjohn: uhm, no... it disables the legacy code (and will enable NGNET in the future" 1134954295 M * derjohn so it's the ngnet bertl is dveloping at the time? I asked myself if there are two kind of "ngnet" 1134954305 A * jayeola build kernel and crosses fingers 1134954356 M * Doener derjohn: yes exactly that ngnet ;) 1134954374 M * derjohn ah, this ngnet! 1134954404 M * derjohn so, the only confusing thing is the description 1134954457 M * Doener yep 1134954586 M * jayeola buiilding the kernel anyway, worst that can happen is that box don't boot :/ 1134954616 M * derjohn Choosing this _disables_ the legacy networking API which is required 1134954616 M * derjohn by the chbind tool. Unless you exactly know what you are doing please 1134954616 M * derjohn leave that option off. 1134954624 M * derjohn would that be clearer? 1134954668 M * jayeola :-) 1134954683 M * jayeola <-- suffering from n00bosis 1134954696 M * derjohn jayeola, I really meant that as question ... 1134954719 M * jayeola "would that be clearer" - yes 1134954888 P * meandtheshell 1134956677 Q * Doener Quit: Leaving 1134958419 J * mep__ mep@p5091B95C.dip0.t-ipconnect.de 1134958859 Q * mep_ Ping timeout: 480 seconds 1134960231 Q * Johnnie Quit: G'bye! 1134962203 J * infowolfe jthm@66-230-102-182-cdsl-rb1.nwc.acsalaska.net 1134962428 J * pflanze ~chris@unk-110.ethz.ch 1134962841 N * Bertl_oO Bertl 1134962845 M * Bertl morning folks! 1134966323 Q * infowolfe Read error: Connection reset by peer 1134966361 J * infowolfe infowolfe@66-230-102-182-cdsl-rb1.nwc.acsalaska.net 1134966361 Q * infowolfe Quit: 1134966554 M * pflanze Hi Bertl 1134966574 M * Bertl hey pflanze! everything fine? 1134966613 M * pflanze yes so far, upgrading to newer kernels have cured the oopses and panics. 1134966631 M * Bertl ahh? what update precisely? 1134966656 M * Bertl (I mean, old version and new version) 1134966706 M * pflanze (strange that I have never seen any oops or panic with the previous kernels for > 1/2 year though, so either it has been some corruption in lvm or so which didn't give a problem anymore with later kernels, or the kernel image itself has been damaged, or I dunno) 1134966722 M * pflanze and I don't run grsec atm. 1134966743 M * pflanze previous: config-2.4.31-vs1.2.10-grsec2.1.5 1134966772 M * pflanze during repair work: vanilla 2.4.32 1134966779 M * pflanze now: 2.6.14.4-vs201 1134966805 M * Bertl okay, so you switched from 2.4-grsec to 2.6 (basically) 1134966809 M * pflanze Now I'm patching grsec into the latter. Hope problems don't return. 1134966827 M * pflanze Yes, and ditched my lvm setup and created a new, lvm2 based one. 1134966829 M * Bertl will be interesting, as there is no grsec for 2.6.14+vserver 1134966867 M * Bertl you know of the 'older' 2.6 grsec + vserver patches? 1134966921 M * pflanze just to remind you, I had a hardware problem (with the disk, I thought), so bought a new disk and copied over the lvm physical volumes and did reiserfsck. But that didn't solve the problems, machine crashed again. After that (but everything on the new disk) I did the kernel upgrade and new lvm setup. 1134966972 M * Bertl yeah, I remember ... 1134966975 M * pflanze yes, I've fetched patch-2.6.11.9-grsec2.1.5-vs2.0rc1-1 from vserver.limehouse.org 1134967000 M * pflanze but I guess it's so old that I'll just do it myself as far as possible 1134967028 M * Bertl yeah, just have a look at what had to be changed to make it work with vserver properly 1134967044 M * Bertl (I consider the patch patches especially interesting here) 1134967069 M * pflanze It will be difficult to be sure. I think I definitely had, or still have, a hardware problem. 1134967124 M * pflanze The effects are twofold: a) the machine just seems to shut down halfway (power LED is still on, but everything else is shut down, no disk or ventilator is running even the one of the power supply). 1134967129 M * pflanze b) oopses and panics. 1134967161 M * pflanze a didn't happen for 3-4 days anymore now; only the oopses+panics. Since my upgrades, those have vanished too. 1134967173 M * Bertl yes, might easily be that the new setup doesn't push your hardware as hard as before 1134967178 M * pflanze I now *hope* a) won't return. 1134967196 M * pflanze (and possibly hose my bits again) 1134967233 M * Bertl of course, a lot of oppses and panics could be the result of broken I/O (at whatever level, disk, bus, filesystem) 1134967265 M * pflanze memtest86 ran fine for a night. I guess it would be rather some bus problem. 1134967296 M * pflanze But I still wonder how it can happen that power simply shuts down. Might be that only 12V power is lost -- ventilators are 12V, the power LED is probably 3.x V. 1134967366 M * pflanze (I suspected a short cut in the disk, since I had many outages in a row, then moved the disk out of the case (so that the disk case didn't touch the machine case) and the outages were gone. 1134967382 M * pflanze But then, with the new disk and 1.5 days later, a new outage happened.) 1134967996 J * malaiwah ~malaiwah@dsl5-019.express.oricom.ca 1134968053 M * Bertl welcome malaiwah! 1134968095 M * malaiwah hi 1134968504 P * malaiwah 1134969578 M * pflanze Hm, in exec.c: 1134969596 M * pflanze standard kernel code: mm->total_vm += vma_pages(mpnt_m); 1134969607 M * pflanze how should vserver code look? 1134969620 M * pflanze (not standard kernel code --- grsec code) 1134969651 M * pflanze vserver introduces vx_vmpages_sub 1134969681 M * Bertl yep, you have to account correctly 1134969684 M * pflanze like vx_vmpages_sub(mm, mm->total_vm - vma_pages(mpnt)); instead of mm->total_vm = vma_pages(mpnt); 1134969695 M * Bertl but I'd suggest you fix the grsec code tjere 1134969712 M * Bertl *there, because the kernel change the way accounting is done in 2.6.14 1134969743 M * Bertl (i.e. they have a central accounting method too) 1134969771 M * pflanze so vx_vmpages_sub(mm, mm->total_vm + vma_pages(mpnt)); i guess 1134969824 M * pflanze no 1134969859 M * Bertl look in older kernels/patches (before 2.6.13 or so, to see how such an 'add' is handled 1134969898 M * pflanze k. I'll finish the trivial stuff first, then come back for the brain stuff. 1134971575 J * balbir ~balbir@59.145.136.1 1134972896 M * Bertl welcome balbir! 1134976352 Q * ryker Quit: Leaving 1134976716 M * Bertl k, off for today .. back tomorrow! 1134976725 M * anonc ah bertl - just wondering if COW should work in 2.1.0-rc10 on amd64? 1134976741 M * Bertl anonc: yes, it's supposed to work there ... 1134976760 M * anonc hmm - I'll try it with x86 and get back to you 1134976788 M * Bertl okay, if you have troubles, enable debugging (especially vserver debugging) 1134976795 M * anonc yep - will do 1134976807 M * Bertl have a good whatever everyone! cya! 1134976812 N * Bertl Bertl_zZ 1134978367 J * _Hunger Hunger.hu@Hunger.hu 1134978459 Q * Hunger Ping timeout: 480 seconds 1134978758 J * Hunger Hunger.hu@213.163.11.138 1134979091 Q * _Hunger Ping timeout: 480 seconds 1134983608 J * Johnnie ~jdlewis@acs-24-154-53-16.zoominternet.net 1134983642 Q * Johnnie Quit: 1134983656 J * Johnnie ~jdlewis@acs-24-154-53-16.zoominternet.net 1134984519 Q * jayeola Ping timeout: 480 seconds 1134985330 Q * pflanze Read error: Connection reset by peer 1134986281 J * pflanze ~chris@unk-111.ethz.ch 1134986718 Q * balbir Quit: Leaving 1134987364 J * Smutje_ ~Smutje@xdsl-84-44-245-30.netcologne.de 1134987470 Q * Smutje Ping timeout: 480 seconds 1134988160 Q * Smutje_ Ping timeout: 480 seconds 1134988231 J * Smutje ~Smutje@xdsl-84-44-245-30.netcologne.de 1134989836 J * _Hunger Hunger.hu@Hunger.hu 1134989884 M * pflanze What should be protected in procfs? 1134989907 M * pflanze is vprocunhide doint the right thing if I don't configure anything? 1134990080 Q * Hunger Ping timeout: 480 seconds 1134991534 N * _Hunger Hunger 1134994745 J * Optik ~optik@carpediem.eintellego.net 1134994777 M * Optik evening all 1134994784 M * Optik anyone available for some assistance? 1134994824 J * dothebart ~willi@xdsl-213-196-221-113.netcologne.de 1134994829 M * dothebart hy... 1134994982 N * Optik Skeeve 1134995555 J * menomc ~amery@200.75.27.52 1134995617 M * _Roey HI 1134995624 M * _Roey HELO 1134995635 M * FaUl HELO ROEY, STOP SHOUTING! :-) 1134995636 M * _Roey http://rafb.net/paste/results/hzGSu076.html 1134995646 M * _Roey FaUl: 840 HELP REQUESTED 1134995664 Q * mnemoc Ping timeout: 480 seconds 1134995664 N * menomc mnemoc 1134995712 M * FaUl _Roey: this is not a vserver-related issue so its likely that you won't get proper help here :-) 1134995730 M * _Roey ? 1134995736 M * _Roey it's channel bonding, come *on*... 1134995782 M * FaUl yes, i see 1134995788 M * JimmyGulp _Roey, pardon my ignorance, but what does channel bonding have to to with vservers? 1134995799 M * _Roey JimmyGulp: failover 1134995804 M * JimmyGulp okay 1134995825 M * _Roey JimmyGulp: because I'm giving each vserver its own IP and I don't want all these services contingent on a single cable. 1134995866 M * JimmyGulp ah, I understand now :) 1134995892 M * FaUl _Roey: have you tried ifconfig eth1 0.0.0.0? 1134995903 M * _Roey hmm no 1134995905 M * FaUl thats how my setup works 1134995909 M * _Roey FaUl: 0.0.0.0? 1134995911 M * _Roey why that? 1134995919 M * _Roey FaUl: OH, I see 1134995920 M * _Roey one sec. 1134995926 M * FaUl because it put the eth0 interface up without any address 1134995944 M * _Roey katzr@sink2:~$ sudo ifconfig eth1 0.0.0.0 1134995950 M * _Roey katzr@sink2:~$ sudo ifenslave bond0 eth1 1134995950 M * _Roey SIOCBONDENSLAVE: Operation not permitted. 1134995965 M * FaUl (i used 802.3ad (mode 3) anyway) 1134995967 M * _Roey Maybe the card itself doesn't support it? 1134995970 M * _Roey ok... 1134995975 M * _Roey FaUl: I put it on mode 0 I think. 1134995993 M * FaUl _Roey: thats what you want if your switch does not support 802.3ad :-) 1134995997 Q * dothebart Read error: Connection reset by peer 1134996012 M * _Roey FaUl: the switch or the card?? 1134996025 J * dothebart ~willi@xdsl-213-196-254-98.netcologne.de 1134996028 M * FaUl _Roey: most cards do, but cheap switches do not 1134996028 M * FaUl :-) 1134996062 M * FaUl hm ,strange 1134996069 M * FaUl i've no system to test right now 1134996094 M * FaUl because the systems that use chan-bonding are on production-work... 1134996137 M * _Roey FaUl: this doesn't work on the machine itself. 1134996143 M * _Roey FaUl: also, this is a Dell Poweredge 1850 1134996153 M * _Roey hardly cheap (or inexpensive either) 1134996193 M * FaUl hm 1134996210 M * FaUl so i don't have any idea, sorry 1134996214 M * _Roey ok 1134996216 M * _Roey thanks though 1134996886 M * dothebart hm, vserver doesn't that much like screen? 1134996894 M * dothebart vserver Willi_Sarge enter 1134996896 M * dothebart mesg: /dev/pts/5: Operation not permitted 1134997018 M * Skeeve hey guys... I was going along really well with the centos howto.. and when ive tried to build a vserver for the first time, all fell in a hole 1134997115 M * Skeeve hmm... bedtime.. ill try again here tomorrow.. 1134997116 M * Skeeve zz 1134997919 Q * matti Ping timeout: 480 seconds 1134998969 M * micah if the default policy of util-vserver is to trust all unknown capabilities, rather than assume unknowns are insecure, I assume this could be a security issue 1134999544 M * micah this was found in util-vserver changelog for 0.30.208 in file lib/getinsecurebcaps.c: 1135000104 J * jeeves ~jeeves@c-24-11-171-10.hsd1.mi.comcast.net 1135001109 J * _Hunger Hunger.hu@Hunger.hu 1135001361 Q * Hunger Ping timeout: 480 seconds 1135001487 J * Hunger Hunger.hu@Hunger.hu 1135001645 J * matti matti@linux.gentoo.pl 1135001831 Q * _Hunger Ping timeout: 480 seconds 1135004436 J * BWare ~bware@office.intouch.net 1135004878 J * Woschak ~Woschak@83-169-156-98-dynip.superkabel.de 1135004896 M * Woschak hello 1135005817 M * TheSeer hi 1135006495 M * pflanze What has to be done to make vserver 2.01 set the hostname of vservers? 1135006512 M * pflanze I read that it needs the new config style, but I'm already using that. 1135006600 M * BWare check /etc/vservers/uts/nodename 1135006607 M * BWare or hostname 1135006626 M * BWare Should all be documented on the flower page 1135006652 M * pflanze yes, I have that file already. 1135006673 M * pflanze I'm using the alpha utils for a long time already, but until now I used vs 1.2 1135006732 M * pflanze Now I've upgraded to kernel 2.6 + vs 2.01, and upgraded the utils from (iirc)0.29.something to 0.30.209 1135006781 M * pflanze and since then, the hostname is not set anymore. 1135007090 M * pflanze The problem happens with those vservers which are started with "plain" init style. 1135007690 M * micah pflanze: when you say it is not set, what do you mean? 1135007708 M * pflanze it is the same as on the host 1135007711 N * Bertl_zZ Bertl 1135007716 M * Bertl good morning folks! 1135007731 M * pflanze hi Bertl 1135007734 M * Bertl (as you can tell from the time, I'm back to Austria :) 1135007775 M * pflanze I've found some entry in old irc logs on the net from a guy with the same problem, CAP_SYS_RAW capability has been the issue then. 1135007792 M * pflanze but, if one has that CAP, why can't I hostname -s myself? 1135007800 M * pflanze (well, maybe ssh removes that CAP?) 1135007801 M * Bertl pflanze: it is very likely _that_ the hostname is set properly ... 1135007818 M * Bertl but also very likely that it is set _twice_ :) 1135007838 M * pflanze but: vserver foo enter , hostname -s foo has no effect 1135007855 M * Bertl how do you verify? 1135007863 M * pflanze entering hostname 1135007883 M * pflanze or cat /proc/sys/kernel/hostname 1135007903 M * pflanze they both still show the host's hostname 1135007916 M * Bertl what does 'grep Cap /proc/self/status' report? 1135007934 M * pflanze CapInh: 0000000000000000 1135007934 M * pflanze CapPrm: 00000000344c04ff 1135007934 M * pflanze CapEff: 00000000344c04ff 1135007936 M * Bertl hey micah! how was your trip? 1135007990 M * Bertl pflanze: I assume testme.sh (v0.14 or later) works fine on the host, yes? 1135008006 M * pflanze ah didn't remember the name of the script 1135008010 M * pflanze going to teest 1135008049 M * pflanze hm where is it? 1135008077 M * pflanze I guess on the web page? 1135008120 M * Bertl google says: http://vserver.13thfloor.at/Stuff/SCRIPT/testme.sh 1135008304 M * pflanze yes, all green. 1135008329 M * pflanze this is without grsec, btw. 1135008347 M * pflanze just vanilla 2.6.14.4+vs201 1135008422 M * Bertl okay, please also upload the /proc/virtual//status and cvirt contents somewhere (e.g. pastebin) 1135008790 M * micah Bertl: its good, very warm here :) 1135008795 M * micah Bertl: welcome home! 1135008812 M * Bertl tx 1135008899 M * micah Bertl: i saw in the changelog of util-vserver 0.30.208 that prior versions trusted all unknown capabilities. In .208 the default policy was changed to assume unknown capabilities are insecure. 1135008924 M * micah Bertl: is there a security problem with by default trusting unknown capabilities? 1135009004 M * micah (trying to evaluate for an update to the debian package) 1135009068 M * pflanze Bertl: http://129.132.126.110/.scratch/out 1135009092 M * pflanze I've added # to the NodeName line 1135009114 M * pflanze the kind of init style doesn't seem related to the problem. 1135009120 M * micah my guess is that trusting unknown capabilities could lead to a host compromise, although I am not certain of this 1135009192 M * micah pflanze: you sure you had uts/nodename set to the vserver hostname before the vserver was started? 1135009202 M * pflanze micah: yes 1135009210 M * pflanze It's an unchanged config. 1135009214 M * pflanze and the file is there. 1135009257 M * micah pflanze: what OS is this? 1135009275 M * pflanze host and guest are both debian sarge 1135009288 M * pflanze as are most of the other guests 1135009302 M * micah pflanze: do you remember the command-line you used to build the vserver guest? 1135009319 M * micah pflanze: also, I assume that /etc/hostname contains the host's name, not the vserver name? 1135009325 M * pflanze the guests are almost two years old 1135009335 M * pflanze I built them manually 1135009369 M * micah you are migrating from legacy 1.2 to 2.? 1135009392 M * pflanze yes 1135009401 M * micah how did you create the files in /etc/vserver? 1135009431 M * pflanze I used the alpha utils already before. 1135009446 M * pflanze Only upgraded kernel patch 1135009454 M * pflanze hey: problem solved. 1135009460 M * micah so your /etc/vserver directory has always been the newstyle configs? 1135009463 M * micah oh? 1135009470 M * pflanze you tipped on it: /etc/hostname had the same string as the host's hostname 1135009506 M * pflanze Now you have to explain me how that string is making it into the hostname of the guest, and hostname -s cannot change it. 1135009525 M * micah what version of util-vserver are you using? Perhaps when you originally built the guest it was a version that did not properly set that 1135009564 M * pflanze I have the old binaries still there; how do I find out their version number? 1135009571 M * Bertl pflanze: VXC_SET_UTSNAME 1135009578 M * mef bertl: glad to see that you made it back safely. 1135009603 M * mef micah: are you still on vacation!? 1135009607 M * Bertl mef: yeah, but what I've learned from that is: never use air france! 1135009636 M * micah mef: well, I am in the tropics, but I have to work still 1135009652 M * micah air france, pouh 1135009660 M * pflanze 0.29.211 1135009661 M * mef bertl: Sorry to hear that you had a bad experience with air france. 1135009677 M * pflanze before I used 0.29.211, now 0.30.209 1135009678 M * Bertl pflanze: it is set in your output, so I ahve no idea _why_ hostname does not work? maybe a debian issue? 1135009686 M * mef bertl: I usually fly continental. maybe next time. ;) 1135009699 M * Bertl mef: we also missed the conenct flight on return, that does tell everything, no? 1135009710 M * pflanze Bertl: it's only the last of the vservers which had the issue, the "elvis" one 1135009730 M * pflanze Bertl: and it was because it had (inside the guest) /etc/hostname set to "elvis" 1135009743 M * Bertl pflanze: yes, that's what I meant with _twice_ 1135009755 M * pflanze with 0.29.211 and vs1.2, that file has been ignored. 1135009772 M * Bertl pflanze: first it is set by the tools, and then by init _inside_ the guest 1135009779 M * mef micah: I bet it sucks having to work from down there, but hey... life is hard! Working from the tropics... that would be my dream! 1135009789 M * mef bertl: sure is telling. 1135009810 M * pflanze Bertl: but why can init set it, but 'vserver 5 enter', 'hostname -s xxx' cannot? 1135009814 M * Bertl mef: but we are here now, so everything is fine again :) 1135009819 M * micah mef: what sucks is that its really hot and the mosquitos have made a feast of me and network connectivity is unreliable, but everything else is great :) 1135009820 M * mef bertl: maybe next time indirect via london using easyjet and then fly one of the many us bound airlines. 1135009827 M * Bertl pflanze: that's what we should figure out ... 1135009837 M * pflanze ok 1135009849 M * mef bertl: or with ryanair to germany somewhere, as there are a ton of flights from there, too. 1135009856 M * Bertl mef: yeah, maybe ... we also have good experiences with united and british ... 1135009899 M * mef bertl: any way, it was super to have you here. Looking forward to the 11 lines of code in the next release that will both give vserver a work conserving fair-share scheduler and disk i/o management! :) 1135009930 M * Bertl shhh, do not spoil everything :) 1135009947 M * mef oh... typo 11,000 lines of code. 1135009947 M * Bertl mef: and yes, it was a great time, thanks for making it possible! 1135009954 M * Bertl LOL 1135009972 M * mef that is: 11,000 lines that wont be necessary or something like that. ;) 1135010047 M * pflanze (Bertl: btw ping now works from inside this guest, traceroute doesn't) 1135010080 M * Bertl yeah, I heard that debian 'fixed' something with traceroute :) 1135010159 Q * shedi Quit: Leaving 1135010492 J * stefani ~stefani@superquan.apl.washington.edu 1135010544 M * Bertl morning stefani! 1135010561 M * Woschak how can i remove a vserver, than i build a new vserver with "vserver build ...." ? 1135010562 M * stefani hi. 1135010570 M * stefani server is down. 1135010580 M * stefani need to go and find out why. 1135010606 M * Bertl Woschak: easiest way is to use --force (which will make a backup of your existing config/guest) which you can remove later ... 1135010630 M * Bertl (--force is an option to vserver build) 1135010827 M * Woschak then i build a vserver with "vserver build --hostname vhost0 ...." i remove it with "vserver build --force vhost0" ? 1135010862 M * Bertl no, the idea was, if there already _is_ a guest named 'karli' 1135010876 M * Bertl and you want to replace it by a new one named 'karli' 1135010878 M * Woschak i start last week with the installation of a vserver and i read, read, read :-) but i don't understand all about vserver at the moment 1135010901 M * Bertl (i.e. remove and rebuild) then the easiest way is to do: 1135010910 M * Bertl vserver karli build --force ... 1135010930 M * Bertl which will 'replace' the old one, and make a copy 1135010934 M * Woschak i see, thank you 1135010971 M * Bertl otherwise, jsut remove config and guest dir 1135010989 M * Bertl (there is no 'delete' or 'remove' command yet) 1135011151 M * micah stefani: did it happen when you stopped a vserver? 1135011155 M * stefani no 1135011171 M * stefani micah: without going there and looking i won't know. 1135011238 M * Woschak Bertl: thank you for the information 1135011248 M * Bertl Woschak: you're welcome! 1135011266 M * Bertl pflanze: could you try out a few things for me on your guest? 1135011492 M * pflanze Bertl: sure -- but could we do it in 1-2 hours? I've an agreement now. 1135011511 M * pflanze BTW, my gentoo user tells me that he cannot reboot his guest. 1135011523 M * pflanze (init style is "gentoo") 1135011579 M * pflanze Bertl: if it's quick, I'll do it now 1135011590 M * Bertl pflanze: yes, 1-2 hours is fine, and the gentoo init style is deprecated, IIRC (ask hollow for details) 1135011606 M * pflanze ok, then see you later. 1135012364 M * Bertl okay, off for dinner now .. back later 1135012372 N * Bertl Bertl_zZ 1135012379 N * Bertl_zZ Bertl_oO 1135014059 Q * click Ping timeout: 480 seconds 1135014361 N * Bertl_oO Bertl 1135014550 M * Bertl back now ... 1135014637 M * pflanze I'm back as well, Bertl 1135014662 M * Bertl pflanze: excellent, need some time or already ready to start? 1135014668 M * pflanze ready to start 1135014669 M * Johnnie Yes, Bertl, you're correct. 1135014682 M * Johnnie The plain initstyle is the only one supported and usable. 1135014725 M * pflanze Johnnie: are you using gentoo? 1135014733 M * Bertl pflanze: okay, let's try: chcontext --xid 100 --secure bash -c "hostname hansi && hostname" 1135014781 M * pflanze New security context is 100 1135014782 M * pflanze hansi 1135014787 J * lilo ~lilo@lilo.usercloak.oftc.net 1135014800 M * Bertl so, that seems like the hostname was set successfully, right? 1135014806 M * pflanze yes 1135014829 M * Bertl let's try the same with the guest like this: 1135014843 M * pflanze root@elvis bin# vserver 5 exec bash -c "hostname hansi && hostname" 1135014843 M * pflanze hansi 1135014856 M * Bertl ah, yes, that's what I had in mind ... 1135014865 M * pflanze ehr 1135014874 M * pflanze what is the -s flag I've always been giving??? 1135014882 M * Bertl don't ask me :) 1135014922 J * click click@ti511110a080-1125.bb.online.no 1135014942 M * Bertl welcome click! 1135014986 M * pflanze ok that was quick. 1135014992 M * Johnnie pflanze: Yes. 1135015021 M * Bertl pflanze: yeah, actually expected something with a little more debugging :) 1135015027 M * click re bertl :/ 1135015037 M * Bertl click: :/ ? 1135015039 M * pflanze The mean thing about it is that as long as /etc/hostname is wrong, "hostname -s fooxyz" simply outputs "elvis" and no error message whatsoever. 1135015049 M * pflanze freaking tools. 1135015052 M * click i'm not much active here 1135015066 M * click keeping track of things reading stuff offline 1135015072 M * click errr, well, backlogs 1135015076 M * Bertl pflanze: hmm, mine say: 1135015078 M * Johnnie I think /etc/hostname is depreciated? 1135015079 M * Bertl hostname -s x 1135015079 M * Bertl hostname: Unknown host 1135015089 M * Bertl (not sure that is better :) 1135015094 M * Johnnie At least on Gentoo anyway... if you're using 2004.3 or 2005.1 or some such. 1135015105 M * pflanze Bertl: yes. It does that too, now that /etc/hostname is "elvis-5". If /etc/hostname is "elvis", it doesn't. 1135015136 M * Bertl okay, anyway for the record, all issues resolved, right? 1135015143 M * pflanze some deep magic inside the hostname algorithm, i guess. 1135015150 M * pflanze yes, everything is fine now. 1135015161 M * Bertl good, that's how it is supposed to be :) 1135015167 M * pflanze Thanks for an ever-improving product! 1135015178 M * Bertl you're welcome! 1135015182 J * lilo_ ~lilo@lilo.usercloak.oftc.net 1135015187 M * Bertl and thanks for helping to improve it! 1135015208 M * pflanze you're welcome =) 1135015301 Q * lilo Ping timeout: 480 seconds 1135016290 M * Woschak i am to looking for a good how to for vserver. the most howtos i found make me confuse. there is no descreption for options and parameter. i compiled a kernel with vserver patch and installed the needed packages (util-vserver). now i want to build a vserver but i have few knowledge about 1135016347 M * Bertl Woschak: have a look at: http://linux-vserver.org/alpha+util-vserver 1135016360 M * Bertl (which explains the tools (somewhat) and gives a few examples 1135016372 M * Bertl and also look at: 1135016421 M * Bertl http://linux-vserver.org/Step-by-Step+Guide+2.6 1135016449 M * Bertl (you basically already did 60% of that, I guess) 1135016484 M * Bertl the configuration itself is described in very detail on the 'flower page' 1135016503 M * Bertl http://www.nongnu.org/util-vserver/doc/conf/configuration.html 1135016535 M * Bertl (make sure to select a css stylesheet which fits your mood) 1135016603 M * Woschak ok, thank you 1135016615 M * Woschak i am in background for a while :) 1135016624 M * Bertl you're welcome! and do not hesitate to ask, if you ahve any questions ... 1135016628 N * Woschak Woschak_read 1135016637 M * Woschak_read ok 1135016758 Q * Johnnie Remote host closed the connection 1135016774 J * Johnnie ~jdlewis@acs-24-154-53-16.zoominternet.net 1135016864 J * lilo ~lilo@lilo.usercloak.oftc.net 1135016891 Q * lilo_ Ping timeout: 480 seconds 1135017446 Q * alexx Ping timeout: 480 seconds 1135017477 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il 1135017484 M * Bertl welcome flock! 1135017513 J * alexx ~alexx@proxy.ikse.net 1135017635 M * flock even though I know it's a script, Bertl, I can't be impolite;) 1135017640 M * flock so hi to you too:) 1135017646 M * Bertl tx :) 1135017688 A * Bertl is now a very happy script :) 1135017689 M * daniel_hozac it's not a script ;) 1135017748 Q * flock Remote host closed the connection 1135017887 J * flock ~restless@l192-117-111-12.broadband.actcom.net.il 1135017904 M * Bertl hey flock! testing the client? 1135017936 M * flock had hdc=noprobe in append="" in my lilo.conf, had to reboot as I added a new disk:) 1135018354 M * jeeves <-- is a scripted bot. My wife has programmed me well. 1135018363 M * Bertl lol 1135018400 M * jeeves jeeves: walk dogs. jeeves: clean kitchen 1135018923 Q * Johnnie Remote host closed the connection 1135018942 J * Johnnie ~jdlewis@acs-24-154-53-16.zoominternet.net 1135018957 Q * Johnnie Remote host closed the connection 1135019317 M * Woschak_read what mee 1135019322 M * Woschak_read sorry 1135019345 M * Woschak_read what mean the --context option at the command vserver? 1135019375 M * Bertl you should assign an unique context id (number between 2 and 49151) for each guest 1135019385 M * Bertl this option is used to specify that id ... 1135019450 M * Woschak_read thx 1135019861 J * Johnnie ~jdlewis@acs-24-154-53-16.zoominternet.net 1135019945 Q * ddlp Remote host closed the connection 1135020976 M * Woschak_read Bertl: thx for this howto (http://linux-vserver.org/Step-by-Step+Guide+2.6). it is very good :) 1135021024 M * Bertl do not thank me, it was written by folks hanging around here ... 1135021031 M * Bertl (see topic :) 1135021037 M * Woschak_read i got the link from you :) 1135021065 M * Woschak_read ok, i thanks the writers of this how to :) 1135021083 M * Bertl yeah, guess they will appreciate it as I do :) 1135021167 M * Woschak_read now i must change my internet contection to the vserver pc, because i want download the new kernel and patch. 1135021186 M * Bertl hmm, why not masquerade it? 1135021198 M * Woschak_read i say good bye for a while :-) 1135021209 M * Woschak_read i maked a new installation 1135021263 M * Woschak_read and i would like to install a vserver in which i install masquerade 1135021296 M * daniel_hozac masquerade is done with iptables, which needs to be run on the host. 1135021307 M * Woschak_read oh 1135021328 M * Woschak_read hm... 1135021339 M * Woschak_read ok, a new question :) 1135021375 M * Woschak_read i would like install a vserver as gateway with squid and iptables for masquerading and nat 1135021404 M * Woschak_read can i make it so? 1135021472 M * Woschak_read the server has two network controller 1135021483 M * Bertl yes, you can do that 1135021499 M * Bertl but you have to configure the masquerading/nat-ing on the host 1135021528 M * Woschak_read only on the host? 1135021566 M * Bertl you can put the squid and whatever inside the guest 1135021578 M * Woschak_read can i bind the vserver with squid and ipdatbles on the outgoing interface only? 1135021587 M * Bertl that's no problem, but networking (with the legacy network) happens on the host 1135021600 M * Woschak_read aha 1135021679 M * Woschak_read ok, last question for the moment :) 1135021695 M * Woschak_read can i install IPCop in a vserver? 1135021707 M * daniel_hozac not likely. 1135021717 M * Bertl yes, but you have to give additional capabilities 1135021734 M * Bertl (and/or) avoid the network restriction 1135021759 M * daniel_hozac pretty much defeating the point of having it in a vserver in the first place ;) 1135021788 M * Bertl well, depends, you get the process and namespace isolation 1135022032 M * daniel_hozac true. 1135022132 M * Bertl Woschak_read: okay, let me explain it to you ... 1135022147 M * Bertl Woschak_read: you know what chroot does? 1135022153 M * Woschak_read yes 1135022170 M * Bertl okay, now there is chcontext and chbind ... 1135022190 M * Bertl and chcontext does what chroot does but for the process space 1135022201 M * Bertl and chbind does the same for the network IPs 1135022223 M * Woschak_read aha 1135022243 M * Bertl for example, if you plan to run soemthing like a firewall 1135022251 M * Woschak_read and have you a manual for this, too? 1135022258 M * Bertl than that is quite fine, but you have to run it on the host 1135022281 M * Bertl yes, there is a paper which describes the virtualizations 1135022328 M * Bertl http://linux-vserver.org/Linux-VServer-Paper 1135022535 M * Woschak_read oh...., many things to read :) 1135022620 Q * Smutje Quit: leaving 1135022793 J * dos000 ~dos000@CPE0013109a586e-CM00137186e53a.cpe.net.cable.rogers.com 1135022799 M * dos000 howdy 1135022805 M * Bertl welcome dos000! 1135022842 M * Woschak_read is it possible that inside a guest a network interface get an ip address from the provider and not the host? 1135022852 M * Woschak_read hello dos000 1135022869 J * Smutje ~Smutje@xdsl-84-44-245-30.netcologne.de 1135022878 M * dos000 anyone know if the vserver-create script is still relevant ? i have version 0.3.4 and i have no idea where i got it from ! 1135022882 M * dos000 hey Woschak_read 1135022884 M * Bertl Woschak_read: not yet (i.e. with legacy networking) 1135022905 M * Bertl dos000: never heard of it, why would you want it? 1135022954 M * dos000 its got debian specific stuff for a clean list of pkg list. but it seems to be based off debootstrap. so it might be old 1135023009 M * Bertl well, vserver - build -m debootstrap uses debootstrap too :) 1135023023 M * Bertl for more than a year now, IIRC :) 1135023052 M * dos000 i have it pasted at http://pastebin.ca/34244 .. just in case 1135023055 M * Woschak_read i want that the host have no directly connection to the internet 1135023154 M * Bertl Woschak_read: won't work 1135023163 M * dos000 the copyright says 2002 ... but i like because of the list of packages that it installs/remove by default 1135023182 M * Bertl dos000: well, you can configure them with util-vserver too 1135023183 M * Woschak_read hm.. 1135023245 M * dos000 Bertl, it also gets rid of the peski xinitd stuff. is it already done with util-vserver as well ? 1135023276 M * Bertl I'm not sure what the xinitd requires 1135023317 M * Woschak_read ok..., i read any more and compile a new kernel 1135023319 J * Vratha ~unix@snow.cs.siue.edu 1135023320 M * Woschak_read cu for a while 1135023323 M * Vratha hello 1135023326 M * Bertl dos000: and a few things in the scrip look useful, but others just look strange 1135023330 M * Bertl welcome Vratha! 1135023334 M * dos000 Bertl, it removes the services that you do not need by default. 1135023340 M * Vratha thanks Bertl :) 1135023420 M * Vratha alright, something odd is happening on my system... i was using vs2.0 on 2.6.8 and using "chcontext --domainname blah.blah.edu vserver blah start", and it worked fine up until an hour or two ago... i can't determine the cause and have tried vs2.0.1 on 2.6.14.3 with no luck... let me paste the message at a pastebin 1135023452 M * Vratha http://rafb.net/paste/results/KjHe1G15.html 1135023480 M * Bertl well, the command line look 'jsut' strange 1135023484 M * Vratha i am confused as to if i set permissions incorrectly somewhere or what?! the vserver i successfully started with that command earlier still have their domainname set 1135023500 M * Bertl no it's like I'd ask: 1135023513 M * Vratha and if i stop and start them using just "vserver blah stop ; vserver blah start", it still has the domainname set 1135023531 M * Bertl I'm doing 'man ls -la' and it just gives me a strange help 1135023555 M * Vratha Bertl: are you talking to me? 1135023567 M * Bertl Vratha: yes, what do you expect your command line to do? 1135023583 M * Vratha oh.. i expect it to start a new security context for the vserver with that domain name set 1135023597 M * Vratha it has worked before; i don't know what's different about it now 1135023601 M * Bertl why not configure it for the guest then? 1135023617 M * Bertl btw, are you sure you want yellow pages domain names? 1135023627 M * Vratha yes i'm sure; we use it only for usernames 1135023634 M * Vratha we use AD/kerberos for auth 1135023644 M * Bertl okay, so just add it to the guest config then 1135023652 M * Bertl and start the guest with: 1135023658 M * Bertl vserver shares start 1135023688 M * Bertl Vratha: I can explain the error message, which is quite correct 1135023701 M * Vratha oh, weird... i did "domainname blah.test" _inside_ the guest and it worked.... a week ago when i tried that i recall it giving me errors saying that i could not set the domainname from inside the guest 1135023721 M * Bertl you _first_ change into a dynamic context (with the chcontext --domainname ac.siue.edu) which is 49155 1135023737 M * Bertl and then try to start the guest (which has a different context id) 1135023751 M * Bertl of course, the context creation from inside the first context is not permitted 1135023755 M * Vratha oh i see 1135023765 M * Vratha i think i understand now; i don't know why it let me get away with that before 1135023786 M * Bertl if that worked somehow with the previous kernel, you can consider that a bug 1135023809 M * Vratha ok, well that makes sense then; i was quite confused about how it just stopped working all of a sudden 1135023816 M * Bertl but adding the domain to uts/domainname should do what you want 1135023836 M * Vratha not a problem; i can do that 1135023837 M * Bertl (/etc/vservers/shares/uts/domainname 1135023840 Q * Woschak_read Ping timeout: 480 seconds 1135023861 M * Vratha i'll take another look at the vserver whitepaper on configs so i know that... i didn't know what "uts/domainname" actually meant at first 1135023867 M * Vratha i thought it was a kernel param 1135023898 M * Bertl no, just means the tree based guest config 1135023912 M * Bertl (if you are with debian, make sure to use the latest versions) 1135023914 M * Vratha yeah, that makes sense 1135023951 M * Vratha i'm using debian sarge with the latest versions in their repository... though i'm about to manually upgrade util-vserver and the kernel+patches 1135023971 M * Bertl check testing/unstable for up-to-date releases 1135023981 M * Bertl (avoid the sarge packages) 1135023996 M * Vratha i may.. i think i'll build my own debian packages; this system is not something i want testing packages to be on 1135024078 M * Vratha anyway; when i put "blah.blah.siue.edu" (literally) in uts/domainname for the guest and then restart the guest, it is retaining the old domainname.. do i now need to use the chcontext to update it? 1135024120 M * Bertl no, how do you check the domainname? 1135024144 M * Vratha `vserver guest enter` then do `domainname` from within the guest 1135024177 M * Bertl and this does not show blah.blah.siue.edu ? 1135024187 M * Vratha no, it does not 1135024195 M * Bertl ah, you are with the debian packages, right? 1135024209 M * Vratha yes... guessing i should just build the latest packages? 1135024216 M * Bertl please replace them by mainstream before testing ... 1135024230 M * Bertl (including the kernel patches) 1135024241 M * Vratha alright, i'll upgrade all that and make some .debs 1135024371 M * Bertl k, you might want to contact micah or aba who usually do the same 1135024598 Q * monrad Quit: leaving 1135024632 M * Vratha to let them know that a .deb has been created or something? 1135024652 M * Bertl well, that or to get a .deb or whatever from them ... 1135024665 M * Vratha oh, well i already created the .deb just now 1135024669 J * monrad ~mikkel@213083190131.sonofon.dk 1135024680 M * Bertl Vratha: okay, that's fine for me :) 1135024684 M * Vratha ;) 1135024768 Q * monrad Quit: 1135024784 J * monrad ~mikkel@213083190131.sonofon.dk 1135024852 M * Bertl wb monrad! 1135025236 M * dos000 Bertl, whats this vreboot thing ? is that still exist ? 1135025260 M * Bertl nope, it was obsolted about one and a half year ago 1135025587 Q * Vratha Quit: leaving 1135026044 M * pflanze How does one reboot with non-plain init style today? 1135026064 M * pflanze There's still this problem that my gentoo guest can't reboot himself. 1135026090 M * Bertl simple, jsut use reboot -f 1135026105 M * Bertl (which does not try to contact init) 1135026187 M * pflanze Heh: I've now created a vshelper logfile, and it prints: 1135026189 M * pflanze Configured (49156) and actual (49158) xid for vserver '/opt/vserver/etc/vservers/bje' are mismatching; aborting... 1135026203 M * pflanze when I do reboot -f inside the guest named "scrat". 1135026228 M * Bertl in general, you should avoid dynamic contexts (like you do right now) 1135026229 M * pflanze scrat has xid 49158 1135026269 M * Bertl assign a context id (e.g. 42) for scrat, and be done 1135026275 M * pflanze "ok". 1135026345 J * meandtheshell ~itsmethem@85-124-35-31.dynamic.xdsl-line.inode.at 1135026362 M * Bertl welcome meandtheshell! 1135026370 M * meandtheshell Bertl: hi 1135027485 M * pflanze How do you assign static xid's ? integer string into /etc/vserver//context, right? 1135027508 M * pflanze I cannot even shut down 'vserver scrat stop' now. It just hangs forever 1135027515 M * Bertl yep, must be between 2 and 49151 1135027523 P * dothebart bye.... 1135027544 M * Bertl pflanze: what tools do you use? 1135027554 M * pflanze newest alpha 1135027561 M * Bertl 0.30.209 1135027568 M * pflanze yes 1135027570 M * pflanze ++ /bin/rm -f /tmp/vserver-lock.V0vlSs 1135027570 M * pflanze ++ /usr/bin/mkfifo -m600 /tmp/vserver-lock.V0vlSs 1135027570 M * pflanze ++ /opt/vserver/lib/util-vserver/lockfile /var/lock/vserver.optvserveretcvserversscrat.startup /tmp/vserver-lock.V0vlSs 1135027570 M * pflanze ++ /bin/grep -q true /tmp/vserver-lock.V0vlSs 1135027601 M * pflanze here it hangs 1135027630 M * Bertl for some time? 1135027650 M * Bertl (there is a shutdown timeout, after 60 seconds or so, the guest gets force killed) 1135027695 M * pflanze the weird thing is that /tmp/vserver-lock.YTbw3S doesn't exist, not even as fifo 1135027709 M * pflanze well /tmp/vserver-lock.V0vlSs or whatever it is atm. 1135027727 M * Bertl looks like a temporary file to me 1135027743 M * Bertl so it was removed from your namespace 1135027747 M * pflanze It seems to create a temporary file, then removes it, and creates a fifo with the same name. 1135027756 M * pflanze Which it wants to use as lock. 1135027764 M * pflanze But it doesn't even exist (anymore). 1135027778 M * Bertl that's the beauty of namespaces :) 1135027791 M * Bertl you can do similar with any bash script 1135027802 M * Bertl (actually it is a bash script :) 1135027820 M * pflanze hm, 1135027828 M * pflanze so you mean there isn't even a race condition? 1135027847 M * Bertl don't know, but I don't think so ... 1135027860 M * pflanze (Since noone else can see the file anyway?) 1135027896 M * Bertl somebody who would have opened the file _before_ it was removed, would still have access 1135027909 M * pflanze Ok, the timeout was 5 minutes. 1135027926 M * Bertl unusual high, maybe your private defaults? 1135027926 M * pflanze I was simply too impatient. 1135027996 M * pflanze well a grep over /etc/vservers/ didn't yield anything containing 300 1135028046 M * Bertl file is called # sync-timeout 1135028061 M * Bertl default is 30 seconds 1135028081 M * pflanze find /etc/vservers/|grep -i time doesn't yield anything 1135028138 M * Bertl double check that you do not have old scripts lying around and the the vshelper (in proc) is pointing to the 'new' binary 1135028146 M * Bertl (s/binary/script) 1135028197 M * pflanze # cat /proc/sys/kernel/vshelper 1135028197 M * pflanze /opt/vserver/lib/util-vserver/vshelper 1135028222 M * pflanze and that is the new one. I did mv /opt/vserver{,_} before installing the new version of the tools. 1135028233 M * pflanze and did ./configure --prefix=/opt/vserver 1135028308 M * pflanze Another problem, or maybe related: vserver scrat start corrently starts halfway through, but never finishes. 1135028314 M * Bertl okay, anyway, that's something where you should look at the cause, not treat the symptoms ... 1135028319 J * comfrey ~comfrey@peoples.coop 1135028328 M * pflanze That also explains why I never got a login prompt at the console of the server. 1135028364 M * Bertl well, what does not finish? or where does it 'hang'? 1135028381 M * pflanze # vserver scrat start 1135028382 M * pflanze ... 1135028383 M * pflanze * Starting spamd ... [ ok ] * Starting Tomcat ... 1135028392 M * pflanze then it waits forever. 1135028399 M * pflanze Now it has started to output: 1135028404 M * pflanze /lib/rcscripts/sh/rc-daemon.sh: redirection error: cannot duplicate fd: Too many open files 1135028407 M * pflanze endlessly. 1135028418 M * Bertl well, looks like your tomcat has issues 1135028424 A * pflanze remembers that "gentoo" init style isn't supported.... 1135028439 M * Bertl well, gentoo init style is deprecated now 1135028454 M * Bertl you use the plain init style for that (according to hollow) 1135028467 M * pflanze ok I'll try to boot with plain and see if it still reboots in a loop 1135029029 M * pflanze Ok, plain init style now works fine. 1135029041 M * pflanze Thanks. 1135029048 M * Bertl you're welcome! 1135029158 Q * jeeves Quit: Leaving 1135029509 J * Vratha ~unix@snow.cs.siue.edu 1135029511 M * Vratha hello :) 1135029515 M * Bertl wb Vratha! 1135029528 M * Vratha thank you Bertl :) how have you been in the last couple hours? 1135029555 M * Bertl fine, fine, and you? 1135029584 M * Vratha pretty good; got everything with the domainname to work, only to find out that this particular guest isn't getting the IP assigned to it so ifconfig can see it 1135029606 M * Vratha even though /etc/vservers/blah/interfaces/0/{ip,dev} exists 1135029632 M * Bertl well, the emphasis is on 'so ifconfig can see it' 1135029646 M * Vratha right... 1135029653 M * Bertl you must know, that ifconfig was replaced more than 5 years ago 1135029666 M * Bertl so it is not able to figure all the things recent kernels do 1135029691 M * Bertl but luckily there is still backwards compatibility left (via aliases) 1135029704 M * Vratha well how come one of my guests is able to see its IP with ifconfig and the other is not? 1135029717 M * Vratha yes, the other guest gets an alias 1135029733 M * Bertl so, just add an entry 'name' with soemthing in it 1135029737 M * Bertl e.g. 'karli' 1135029763 M * pflanze BTW, where are boot messages written to when using plain init style? 1135029780 M * Bertl pflanze: in the log files (where init puts them) 1135029882 M * pflanze I mean the "foo [ok]" etc. messages, console messages. They are not in /var/log/messages 1135029883 M * Vratha eh, i don't know; i put "test" in /etc/vservers/nis/interfaces/0/name, and restarted the vserver with the same result (e.g. no alias with the IP)... is there something i need to do with chbind first? 1135029943 M * Bertl did you stop the guest before doing the change? 1135029958 M * Bertl if not, then it did not succeed to remove the ip 1135029966 M * Bertl (which you now have to do by hand) 1135030018 M * pflanze BTW it seems that lsof can't show open filehandles of guest init processes? 1135030019 Q * meandtheshell Remote host closed the connection 1135030027 M * pflanze chcontext --ctx 1006 lsof -p 1666 # yields no output 1135030043 M * Bertl pflanze: does that guest use plain init style? 1135030043 J * meandtheshell ~itsmethem@85-124-9-10.dynamic.xdsl-line.inode.at 1135030046 M * pflanze ah. ctx 1. 1135030049 M * Bertl wb meandtheshell! 1135030049 M * pflanze Bertl: yes 1135030056 M * Vratha no, i didn't stop the guest before creating that file; i created it, stopped the guest, and restarted it with the same result... however, i did "chbind --ip vserver nis start", and it got the correct IP alias now 1135030066 M * meandtheshell Bertl: hi ;) 1135030074 M * pflanze ehr, no, same thing with ctx 1. 1135030085 M * Bertl Vratha: stop doing the chcontext/chbind plus guest startup 1135030095 M * Vratha hey, it's what works 1135030103 J * Doener doener@i5387CBF1.versanet.de 1135030112 M * Bertl Vratha: but just by 'accident' 1135030124 M * Vratha well the way you told me didn't work correctly 1135030129 M * Doener evening! 1135030138 M * Vratha i mean just by adding that file and restarting the vserver 1135030148 M * Bertl Vratha: because you did something wrong (we do not know yet) 1135030169 M * Bertl Vratha: but we will figure it ... I'm sure 1135030192 M * Bertl stop the guest again, and restart it, then do: 1135030220 M * Vratha ok, it is restarted, without using chbind or anything like that 1135030329 M * Bertl vserver nis exec cat /proc/self/ninfo 1135030385 M * Vratha ok, that shows my NID, along with the IP address for it on V4Root[0] 1135030396 M * Bertl okay, then let's do: 1135030406 M * Bertl vserver nis exec ip addr ls 1135030417 M * Bertl (if you do not have iprout2 inside the guest do) 1135030430 M * Vratha i do not have iprout2 1135030432 M * Bertl ip addr ls | grep 1135030456 M * Vratha oh, but i have it on the host 1135030465 M * Bertl yeah, that's what I meant :) 1135030481 M * Vratha inet 146.163.150.5/24 brd 146.163.150.255 scope global secondary eth0:0 1135030492 M * Vratha there is also another line dealing with global scope 1135030501 M * Vratha i think a primary 1135030509 M * Bertl with the same IP? 1135030514 M * Vratha yes 1135030518 M * Vratha inet 146.163.150.5/32 scope global eth0 1135030537 M * Bertl so, you managed to set two different networks here 1135030551 M * Bertl one with /32 (host) and one with /24 1135030572 M * Bertl it also seems as if your guest gets the same IP assigned as your host? 1135030588 M * Bertl (or where does the 146.163.150.5/32 come from? 1135030598 M * Vratha 146.163.150.5/32 is an alias on the host 1135030607 M * Vratha i put them there so the guests could use them 1135030615 M * Vratha was that incorrect? 1135030618 M * Bertl so, you have to decide what you want 1135030626 M * Vratha i want the guest to use them 1135030629 M * Bertl you ahve three options basically 1135030645 M * Bertl a) you assign the IP on the host, and just let the guest use it 1135030658 M * Bertl you do that by specifying ip and nodev 1135030674 M * Bertl b) you let the tools assign the IP for the guest, new-style 1135030692 M * Bertl just specify dev, ip and prefix (or netmask) 1135030708 M * Bertl c) the tools should create an alias (legacy) 1135030719 M * Vratha oh, in b's case, i would just remove the alias from the host and let "vserver nis start" take over? 1135030721 M * Bertl specify dev, ip, prefix (or netmask) and name 1135030733 M * Bertl yes, precisely 1135030809 M * Vratha sweet; that sounds pretty simple 1135030849 M * Bertl it actually is, you can even decide that at creation time 1135030861 M * Vratha where exactly are the "good docs" for vserver? none of the previous docs i found mentioned any of this 1135030881 M * Bertl http://linux-vserver.org 1135030887 M * Vratha yeah i'm there 1135030888 M * Bertl http://linux-vserver.org/Linux-VServer-Paper (theory) 1135030906 M * Vratha oh, right.. i haven't finished reading that one yet 1135030909 M * Bertl http://www.nongnu.org/util-vserver/doc/conf/configuration.html 1135030917 M * Bertl (the configuration description) 1135030930 M * Bertl http://linux-vserver.org/Step-by-Step+Guide+2.6 1135030935 M * Bertl (the start guide) 1135030966 M * Vratha yeah, i've read all those except the fine details of the config page and the rest of the vserver-paper 1135030985 M * Vratha barely any of vserver-paper so far; that will probably answer more of my questions 1135030998 M * Bertl and of course, always feel free to extend/improve the docu 1135031024 M * Bertl (which is probably necessary) 1135031081 M * pflanze Should I write two mails to the list with these questions?: 1135031093 M * Vratha thanks for your help so far Bertl :) 1135031098 M * pflanze - lsof on a guest init doesn't work 1135031113 M * pflanze - where does console output go with plain init style? 1135031136 M * pflanze (It's not urgent, and so might help others as well.) 1135031157 M * Doener pflanze: hm, what does the first question mean? 1135031164 M * Bertl Vratha: you're welcome! 1135031184 M * pflanze Doener: 1135031186 M * pflanze # vps aux|grep scrat|grep init 1135031186 M * pflanze root 1666 1006 scrat 0.0 0.0 1464 496 ? S lsof -p 1 1135031255 M * pflanze Bertl: root@elvis root# chcontext --ctx 1006 lsof -p 1666 1135031255 M * pflanze lsof: WARNING: can't stat() usbfs file system /proc/bus/usb 1135031255 M * pflanze Output information may be incomplete. 1135031259 M * pflanze same thing. 1135031283 M * Bertl Vratha: yes, the 'name' is the one which is required for the alias (see c) 1135031297 M * pflanze (lsof -p 1 on the host works) 1135031315 M * Doener pflanze: -p 1, not -p 1666 when using --ctx 1006 1135031322 M * pflanze ah 1135031336 M * pflanze funny :)) 1135031339 M * pflanze thanks. 1135031380 A * Bertl .o( guess if folks would read what I write a lot of things would be much easier :) 1135031428 M * Vratha Bertl: well, i created "name", and when i enter the guest and do 'ifconfig', i see the alias, but i do not see the IP... so in interfaces/0 i have dev (eth0), ip (146.163.150.5), prefix (24), and name (0) 1135031468 M * pflanze And the second question? (where does "console output" of the boot process go?) 1135031472 M * Bertl Vratha: check again, with 'ip addr ls' that you do not have two entries 1135031481 M * Vratha i do not have two entries 1135031485 M * Bertl Vratha: also, what does your ifconfig show? 1135031511 M * Vratha also, when i shut down the vserver, the only entry that i had in "ip addr ls" for 146.163.150.5 disappears 1135031512 M * Bertl pflanze: read up, I already answered that one :) 1135031520 M * Vratha ifconfig on the host does not have that IP 1135031521 M * pflanze ah sorry 1135031556 M * Vratha Bertl: i think i found the cause 1135031584 M * pflanze Bertl: well you said, to the init logs. But which ones? /var/log/{messages,syslog} don't contain this output. 1135031593 M * Vratha Bertl: my host has 3 other aliases for other purposes, numbered 0,1,2, and in name i had "0"... i think since i tried using alias 0 when the same alias on the host was using a different IP caused the problem 1135031610 M * Bertl pflanze: did you check _inside_ the guest? 1135031626 M * pflanze Bertl: yes 1135031638 M * Bertl Vratha: well, looks like :) 1135031667 M * pflanze Bertl: only those messages that daemons and the like write to syslog are being put there, which makes sense. Usually the output being written to the console isn't logged. 1135031820 J * Akito ~akito@67.71.244.170 1135031845 M * Doener pflanze: # /etc/vservers/.defaults/apps/init 1135031845 M * Doener * tty 1135031845 M * Doener A symlink to the TTY device where input/output will be redirected from/to at startup via initscript. 1135031867 M * Bertl welcome Akito! 1135031872 M * Akito Hello =) 1135031886 M * Akito New to irc, somewhat, so be gentle ;-) 1135031892 M * Doener pflanze: (same is available on a per vserver basis) 1135031912 M * Doener Akito: Bertl always is ;) 1135031929 M * Akito He he. Nice to know beforehand. 1135031943 M * Akito I was wondering if anyone could help me with something. 1135031968 M * Doener if you don't mind telling us what the problem is ;) 1135031978 M * Bertl Akito: I'm pretty sure someone can help you with anything :) 1135031982 M * Akito I'm playing around with vservers enough lately, and despite the security risks, I need 1 specific vserver to have access to the cd burner. 1135032013 M * Bertl okay, first step, copy over the required devices to the guest's dev 1135032019 M * Akito Not being too used to creating devices, I managed to find info to create the /dev/hdc device in my vserver file system. 1135032061 M * Vratha Bertl: sweet; thanks for the help! it's all working now 1135032076 M * Bertl Akito: just copy 'cp -va' them from the host 1135032077 M * Akito So basically, copying the /dev/hdc to say /vservers/server1/dev would be enough for step 1? 1135032088 M * Akito mmmkay, doing that now. 1135032095 J * Smutje_ ~Smutje@xdsl-87-78-85-173.netcologne.de 1135032096 M * Doener Akito: yes. and cp -a does the job, as you already have the dev node on the host, there's no need to mknod it ;) (though it can't hurt to know how to do it) 1135032205 Q * Smutje Ping timeout: 480 seconds 1135032206 M * Akito other apps should be able to access the device after the vserver is restarted, I assume? 1135032405 M * Doener Akito: no restart needed 1135032439 M * Doener what might be needed are additional capabilities (which would then require a restart) 1135032454 M * Akito Starting to think my problem may lie elsewhere 1135032477 M * Akito hmmm, has anyone ever heard of a cd-burner being able to be used from a vserver? 1135032477 M * Bertl what problem do you ahve? 1135032494 M * Bertl doesn't it give any messages? 1135032521 M * Akito well, basically, not many. I'm trying to go as low level as I can and start from there. 1135032532 M * Akito Right now, trying to use cdrecord to see if it can even see the device 1135032549 M * Doener Akito: you most probably need additional capabilties. If there's some error message, please show it to us. otherwise stracing the failing process should give some hint about the missing capability. 1135032735 M * Akito Thanks for the help so far =) I'm gonna try and get familiar with strace a bit more, to isolate where my problem should be. Gonna start by making sure I can access the burner from the host first, and start from there. I just didn't find any info on if it was even possible at first. 1135032788 M * Doener Akito: for stracing the process, just use: strace -fF , e.g. strace -fF ps -Af 1135032799 M * pflanze Doener: this doesn't seem to work. 1135032814 M * Doener (if you want to share the strace, please use pastebin.com or some other pastebin service) 1135032835 M * Akito he he, sure, didn't intend to flood the channel. 1135032838 M * pflanze Doener: (setting apps/init/tty has no effect) 1135032852 M * Akito but I think my work day will be over before I get to that point. 1135032876 A * pflanze wonders what 'initscript' is 1135033090 M * pflanze Doener: actually, the apps/init/tty file is truncated upon vserver restart! 1135033099 M * pflanze Looks like a bug. 1135033152 M * Doener pflanze: hu? how do you truncate a symlink? 1135033154 M * pflanze ehr. I did echo /dev/vc/7 > apps/init/tty instead of symlinking 1135033162 M * pflanze hehe 1135033227 M * pflanze ok, now did the symlink. 1135033250 M * pflanze 'echo Hello > apps/init/tty' outputs something on the console 7 1135033269 M * pflanze but rebooting the respecting vserver (which does have init style plain) does not output anything. 1135033419 M * pflanze I even did cp -a /dev/vc /vservers/5/dev/ just to see whether it means to use the devices inside the guest, but that doesn't help either. 1135033458 M * Bertl pflanze: how aboud doing: 1135033462 M * Bertl *about 1135033481 M * Bertl cp -a /dev/vc/7 /vservers/5/dev/console ? 1135033492 M * pflanze good idea 1135033552 M * pflanze yeah! that outputs stuff there. 1135033578 M * mnemoc Bertl: what happened with the idea of using unix sockets there? 1135033588 M * Bertl mnemoc: lost in vision :) 1135033593 M * mnemoc :) 1135033632 M * pflanze Does info about this bug somehow make it to Enrico, or whomever need to know about it, or do I have to send a mail? 1135033645 M * Akito Have a nice evening, everyone! Will continue this tomorrow, need to focus on a problem at work first. Thanks for the welcome, and help so far! 1135033648 M * Bertl don't see a bug yet 1135033661 M * pflanze Bertl: apps/init/tty doesn't work as advertised. 1135033667 M * Bertl Akito: you're welcome and feel free to hang around! 1135033676 M * Bertl pflanze: who did advertize it as what? 1135033678 M * Doener pflanze: i just checked, my fault i guess 1135033683 M * pflanze Bertl: the flower page 1135033695 Q * Akito Remote host closed the connection 1135033697 M * Bertl pflanze: gimme facts :) 1135033713 M * Doener pflanze: it is just supposed to redirect output from the "vserver" script, f.e. when called from the initscript 1135033717 M * pflanze Bertl: read Doener's advice to me above. 1135033734 M * Bertl which was a little to fast ... 1135033787 M * Bertl # 1135033787 M * Bertl * 1135033787 M * Bertl A symlink to the TTY device where input/output will be redirected from/to at startup via initscript. 1135033794 M * Doener the apps/init/tty stuff is used for vserver (the script) output when used like: vserver --defaulttty foo start 1135033834 M * Bertl that doesn't mean that it will somehow (magical) manage to redirect the console for init 1135033842 M * pflanze Doener: hm, but that output is still written to the vshelper log 1135033874 M * pflanze "that" meaning the output of the vshelper, sorry. 1135033908 M * pflanze ok, I think I can live with /dev/console inside the guest. 1135033931 M * pflanze I'll now have to find a way to preserve stuff written to it over reboot, so that the guest user can see the output afterwards. 1135033950 M * Bertl that's why 'init' usually logs to the syslog 1135034015 M * pflanze hm, but you have to configure it to do so? 1135034027 M * pflanze since by default (on debian and gentoo) it doesn't. 1135034106 M * Bertl pflanze: hmm, don't know, mandrake does that, for example 1135034319 Q * flock Ping timeout: 480 seconds 1135035266 Q * Vratha Quit: leaving 1135035516 P * meandtheshell 1135036176 P * stefani I'm Parting (the water) 1135036243 Q * pflanze Quit: good night