1341274985 P * chas 1341280450 J * chas ~charlie10@82-69-206-182.dsl.in-addr.zen.co.uk 1341280814 M * chas Hi all, a while ago I was asking: how to compile a 64bit kernel with vserver patch on a debian mixed architecture. 1341280823 M * Bertl yup 1341280824 M * chas Ben kindly helped me out. 1341280842 M * chas My question now is: Once I create the chroot environment and build the kernel how do I install it? 1341280859 M * chas I thinking here that I would have two boot directories one in the host env and one in the chroot env. 1341280868 M * chas Do I have to mount the host /boot inside the chroot env? 1341280882 M * Bertl do 'make help', look for a debian related package :) 1341280974 M * chas Sorry is simple but I don't understand what's being said here? 1341280993 M * chas I keep missing words out. 1341281019 M * chas The command is simple enough but I don't understand what is intended? 1341281079 M * Bertl okay, enter the kernel build tree, do: 'make help | grep deb' 1341281275 M * chas Ok I see, but which env do I run the installation in? I'm assuming I run it in my chroot env as the host/mixed gives me errors. 1341281294 M * Bertl you run the make in the chroot, you get a package 1341281308 M * Bertl you should be able to install that on the host 1341281328 M * Bertl if not, you can still do a make modules_install and make install 1341281344 M * Bertl and simply copy over the /lib*/modules/* and /boot/vm* 1341281375 M * chas OK that sounds like a good work around. 1341281401 M * Bertl but double check first with the package, it should do the trick 1341281446 M * chas I'm asking because I managed to download this package: linux-image-2.6.32-5-vserver-amd64_2.6.32-45_amd64.deb and I get error trying it in my host env. 1341281495 M * chas It's late here not but I will try the chroot in the morning. Thank again for your help. 1341281513 M * Bertl np, have fun (or a good sleep :) 1341281534 P * chas 1341285726 Q * clopez Ping timeout: 480 seconds 1341291404 J * ghislain ~AQUEOS@adsl2.aqueos.com 1341293689 M * uranus Bertl, is there any chance to debug such dmesg outputs? http://paste.linux-vserver.org/22861 1341294046 M * Bertl well, yes, to some degree 1341294059 M * Bertl what kernel/patch is that? 1341294084 M * uranus 3.2.21 with 2.3.2.10 1341294207 M * Bertl well, the first step is to find all occurances of this byte sequence in the kernel code 1341294246 M * uranus simple grep` 1341294247 M * uranus ? 1341294248 M * Bertl once that has been found, you can map the place(s) to functions and specific object files (given that you have the kernel build tree or can recreate it) 1341294264 M * uranus build tree is there 1341294291 M * Bertl grep will fail at the non ascii/special bytes 1341294302 M * Bertl i.e. you want a 'binary grep' for that 1341294310 M * uranus thx 1341294332 M * Bertl assuming that there are not that many occurances, you might simply use hexedit to search for the sequence 1341294370 M * uranus i hoped to find in 3.2 a more stable version than 3.3, because in 3.3 we have after a few days secure-mount processes in state d and also kswapd in state d 1341294398 M * Bertl did you try with 3.4 yet? 1341294412 M * uranus no that's the next step for this day 1341294464 M * Bertl btw, interesting regarding secure mount, that sounds like you are doing many stop/start cycles in your setup? 1341294472 M * uranus yes 1341294501 M * Bertl so it might be easily triggerable with a start/stop or even restart loop? 1341294512 M * uranus maybe some off the restarts are under heavy I/O, but disabling blkio did not solve that issue 1341294518 M * uranus maybe 1341294528 M * uranus good point 1341294534 M * Bertl for example, have a bunch of guests (like 5-10) 1341294545 M * Bertl and restart them in parallel in a loop 1341294553 M * Bertl (i.e. one loop for each of them) 1341294574 M * uranus just a moment, will setup the loop 1341294594 M * Bertl if that doesn't shake out something (if there is anything related to this setup) ... 1341294702 M * Bertl btw, there seems to be a dnotify/inotify related issue on kernels up to 3.4 which causes strange and unexpected delays 1341294720 M * uranus 3.4 is affected also? 1341294733 M * Bertl it was fixed in 3.4 IIRC 1341294747 M * uranus Bertl, http://paste.linux-vserver.org/22862 should do the trick or? 1341294762 M * uranus oh well, that sounds nice :), have to test 3.4 after the loop 1341294781 M * Bertl do you usually issue restarts? 1341294799 M * uranus hmm no 1341294802 M * uranus stop and start 1341294805 M * Bertl if yes, then the loop should be fine, can be written tighter like this though: 1341294818 M * Bertl while true; do vserver restart; done 1341294831 M * Bertl if you use stop and start, then I'd use that as well 1341294869 M * Bertl something like: while true; do vserver start; sleep 1; vserver stop; sleep 1; done 1341294879 M * uranus set -e should brake my loop right? 1341295004 M * Bertl IMHO it has no effect in your script 1341295011 M * uranus the import thing at this state is to get a secure-mount in state d to check reproduceability right? 1341295047 M * Bertl yes, although a kernel oops/trace could be nice too :) 1341295094 M * uranus i did not get any of them 1341295211 J * ncopa ~ncopa@3.203.202.84.customer.cdi.no 1341295715 J * FireEgl ~FireEgl@173-16-9-169.client.mchsi.com 1341296471 M * uranus Bertl 5 restarts running now in parallel more will follow, till now no secure-mount in state D 1341296994 Q * ncopa Quit: Leaving 1341297279 J * ncopa ~ncopa@3.203.202.84.customer.cdi.no 1341300484 Q * FireEgl Quit: Leaving... 1341302843 Q * uranus Quit: Verlassend 1341302856 J * uranus ~uranus@62.152.161.117 1341304535 J * FireEgl FireEgl@2001:470:e5ad:1:39bc:7c86:c6bf:ea69 1341304849 J * plastics ~plastics@c-69-138-42-222.hsd1.tn.comcast.net 1341305195 J * petzsch ~markus@dslb-092-078-112-205.pools.arcor-ip.net 1341305336 Q * plastics Ping timeout: 480 seconds 1341307364 Q * brambles Quit: leaving 1341307401 J * brambles brambles@79.133.200.49 1341307465 J * fisted_ ~fisted@xdsl-84-44-145-48.netcologne.de 1341307754 Q * fisted Ping timeout: 480 seconds 1341309116 Q * petzsch Quit: Leaving. 1341310274 J * clopez ~clopez@fanzine.igalia.com 1341312700 J * petzsch ~markus@dslb-092-078-112-205.pools.arcor-ip.net 1341312910 J * BenG ~bengreen@cpc2-oxfd11-0-0-cust221.4-3.cable.virginmedia.com 1341314348 M * Bertl off to bed now ... have a good one everyone! 1341314352 N * Bertl Bertl_zZ 1341314978 Q * petzsch Read error: Connection reset by peer 1341317936 J * plastics ~plastics@c-69-138-42-222.hsd1.tn.comcast.net 1341318865 Q * Aiken Remote host closed the connection 1341322831 Q * plastics Ping timeout: 480 seconds 1341323094 J * DreamerC ~DreamerC@122-116-181-118.HINET-IP.hinet.net 1341324807 J * plastics ~plastics@c-69-138-42-222.hsd1.tn.comcast.net 1341326325 Q * BenG Quit: I Leave 1341327446 M * ghislain if i want to check a guest with a host binary ( to prevent rootkit modified one) how can i do this ? if i use vnamespace -e xx /bin/ps it will use the host binary in the guest context, will this work ? 1341327482 N * ensc Guest2020 1341327492 J * ensc ~irc-ensc@p54ADFBB0.dip.t-dialin.net 1341327877 Q * Guest2020 Ping timeout: 480 seconds 1341331335 J * BenG ~bengreen@cpc10-aztw24-2-0-cust114.aztw.cable.virginmedia.com 1341331376 Q * hparker Ping timeout: 480 seconds 1341331828 J * bonbons ~bonbons@2001:960:7ab:0:89a3:30ab:7575:7b8c 1341332808 N * Bertl_zZ Bertl 1341332813 M * Bertl morning folks! 1341332951 M * uranus moin Bertl my restart loop did not provide any secure-mount in State D after 10 hour 1341333131 Q * plastics Ping timeout: 480 seconds 1341333607 M * Bertl okay, so it is unlikely that it is start/stop related 1341333626 J * bergerx ~bergerx@2-228-78-73.ip190.fastwebnet.it 1341333648 M * Bertl i.e. I suspect it might be either coincidence or requires some other yet unknown ingredient 1341334867 Q * ensc|w Remote host closed the connection 1341334876 J * ensc|w ~ensc@www.sigma-chemnitz.de 1341337221 Q * fisted_ Ping timeout: 480 seconds 1341339240 J * fisted ~fisted@xdsl-84-44-145-48.netcologne.de 1341339327 Q * BenG Quit: I Leave 1341339898 Q * bergerx Ping timeout: 480 seconds 1341342493 J * bergerx ~bergerx@2-228-78-73.ip190.fastwebnet.it 1341342802 J * plastics ~plastics@c-69-138-42-222.hsd1.tn.comcast.net 1341346348 J * patulo c8447022@ircip2.mibbit.com 1341346411 M * patulo Hi, I need to know if it's possible to run virtualbox inside a vserver guest with this kernel 2.6.32-5-vserver-amd64 1341346414 M * patulo any ideas ? 1341346419 M * patulo thanks ! 1341346483 M * Bertl no idea, probably virtualbox will make use of the kvm kernel interface or similar 1341346501 M * patulo it asks for dkms 1341346503 M * Bertl in which case you might need some extra capabilities to get it work 1341346523 M * Bertl s/it work/it to work/ 1341346526 M * patulo could u tell me wich capabilities should I ? 1341346537 M * patulo which 1341346538 M * patulo :P 1341346542 M * Bertl depends on where it fails, strace -fF should help 1341346713 M * patulo ok, I'm stracing it now 1341346835 M * patulo Could not load the Host USB Proxy service: VERR_NOT_FOUND. 1341346849 M * patulo that error I suppose is cause it want access to the USB hub 1341346856 M * patulo wich actually I don't really need 1341346951 M * patulo the red message that appears in the Vbox says "The value of the Processor(s) field on the System page is not complete." 1341346962 M * Bertl well, then disable it (the usb part) 1341346965 M * patulo but it doesn't crash 1341347005 M * patulo I just disabled it now 1341347030 M * patulo the guest is running with the standard capabilities 1341347072 M * patulo it's like it lacks of processors 1341347174 N * BobR BobR_zZ 1341347441 M * patulo i just add the VIRT_CPU to the ccapabilities and the guest fails to start with this msg... "Unknown ccap 'VIRT_CPU'" 1341347478 M * patulo do u know what I'm doing wrong ? 1341347597 M * Bertl VIRT_CPU is a cflag not a ccap, so yes, I know what you are doing wrong :) 1341347632 M * patulo jeee :P 1341347633 M * patulo ok !! 1341347634 M * Bertl but VIRT_CPU is not very likely to help you with virutalbox 1341347636 M * patulo thnx ! 1341347641 M * Bertl what does the strace say? 1341347655 M * patulo nothing, cause it doesn't fails actually.... 1341347667 M * patulo or too much... ! 1341347683 M * Bertl so everything fine with virtualbox? 1341347706 M * patulo the vbox interface starts 1341347714 M * patulo but I can start a VM 1341347716 M * patulo cant 1341347725 M * patulo it can't see the processor 1341347757 M * Bertl and what does "can't see" mean? 1341347873 M * patulo in the screen where you can set how many processors you can assign to a VM it shows 0 processors available 1341347891 M * patulo good question, by the way :P 1341347909 M * Bertl so it probably got that info somewhere (or rather didn't get it) 1341347928 M * Bertl check with the strace output where it does do the "looking" :) 1341347941 M * patulo jeee 1341347943 M * patulo yeah :P 1341347986 M * patulo ooops 1341348004 Q * clopez Ping timeout: 480 seconds 1341348017 M * patulo I don't know what happend, I just removed the capabilities and now it doesn't start !.... 1341348026 M * patulo root@vpsdesa:/etc/vservers/x# rm bcapabilities.no ccapabilities.no root@vpsdesa:/etc/vservers/x# vserver x start chroot-sh: open("/etc/mtab"): Read-only file system chroot-sh: unlink("var/run/motd"): Read-only file system chroot-sh: unlink("var/run/crond.reboot"): Read-only file system 1341348090 M * patulo this server it's playing with me... 1341348166 M * Bertl be careful, servers can be vicious sometimes 1341348218 M * patulo I rebooted and the guest starts ... :P 1341348230 M * patulo maybe is a Windows :P 1341348356 M * Bertl yeah, recent distros try to immitate Windoof or MacOX behaviour very closely 1341348381 M * patulo it's a squeeze !!!! not Mint/ubuntu :P ! 1341348386 M * patulo It should be decent ! 1341348636 Q * sannes Remote host closed the connection 1341348653 M * patulo is there any way to give all privileges to a guest ? 1341348662 M * patulo its a trusted environment 1341348710 M * Bertl yes, you can give all capabilities (bcapabilities) to the guest 1341348735 M * patulo I'll try that just to be sure that is a matter of caps ;) 1341348740 M * patulo or not ... 1341348747 M * Bertl but if it is looking in /proc somewhere, you might need to unhide the entries, and if it uses a device, you might want to add that one 1341348787 M * patulo I'll give it another look at the strace 1341349087 J * Aiken ~Aiken@2001:44b8:2168:1000:21f:d0ff:fed6:d63f 1341349171 Q * plastics Ping timeout: 480 seconds 1341349380 Q * bonbons Quit: Leaving 1341349658 M * patulo it's not a matter of bcaps !! jeeee 1341349665 M * patulo it fails anyway 1341349678 M * patulo Im now trying to strace it 1341349699 M * patulo to catch when it tries to "see" the processors 1341349988 M * patulo does it mean anything to you: "[pid 2413] read(7, 0x7e77f4, 4096) = -1 EAGAIN (Resource temporarily unavailable)" 1341349991 M * patulo EAGAIN ? 1341350010 M * Bertl yes, it means the system call was interrupted 1341350017 M * Bertl (and the process should try again) 1341350034 M * patulo should I look before that, shouldn't ? 1341350058 M * Bertl that happens when a signal is received for example, so that's not really unusual 1341350068 M * patulo ok 1341350081 M * patulo the nohup file I got is 445000 lines :P 1341350116 Q * bergerx Ping timeout: 480 seconds 1341350346 M * patulo [pid 2423] access("/etc/ld.so.nohwcap", F_OK [pid 2422] <... brk resumed> ) = 0x1406000 [pid 2423] <... access resumed> ) = -1 ENOENT (No such file or directory) [pid 2422] access("/etc/ld.so.nohwcap", F_OK 1341350346 M * patulo what about that ? 1341350497 M * Bertl that's normal unless you want to deactivate hardware optimized versions of system libraries 1341350508 M * patulo ok, tnhx ! 1341350669 M * Bertl you might also take a peek at the source code, maybe something obvious stands out 1341350680 M * patulo yeah 1341350694 M * patulo I'm doing that right now.... 1341350997 J * clopez ~clopez@131.29.165.83.dynamic.mundo-r.com 1341351152 M * patulo could you guess where in the /proc can the Vbox look for procinfo... I just know /proc/pcuinfo and /proc/stat... 1341351162 M * patulo but there is no error looking for that in the strace 1341351270 M * Bertl no idea, but I'd suggest to search the sources for the errormessage 1341351279 M * Bertl (you get an error message, yes?) 1341351293 M * patulo "The value of the Processor(s) field on the System page is not complete." 1341351316 M * patulo that... but that string doesnt exist in my nohup ! 1341351386 M * Bertl search in the vbox source 1341351554 J * fisted_ ~fisted@xdsl-87-78-17-177.netcologne.de 1341351838 Q * fisted Ping timeout: 480 seconds 1341353868 Q * ghislain Quit: Leaving. 1341354212 M * nkukard this is rather odd, I'm getting a setruid -EPERM error and i've assigned CAP_SETUID to the guest in bcapabilities 1341354215 M * nkukard I must be missing something 1341354240 M * nkukard setreuid 1341354498 M * nkukard hrmmm 1341354553 M * Bertl CAP_SETUID is given by default 1341354562 M * nkukard exactly, I just saw that too 1341354583 M * nkukard setreuid(0, 0) = -1 EPERM (Operation not permitted) <= getting that in a setuid binary 1341354597 M * Bertl kernel/patch? 1341354622 M * nkukard 2.6.32.25, patch , one sec 1341354757 M * nkukard Bertl, I'm unable to determine the patch, is there a way I can do so from within the system itself without having the sources? 1341354913 M * Bertl not really 1341354933 M * Bertl the patch version is usually part of the kernel version name 1341354946 M * Bertl (unless a distro maintainer removes it :) 1341354949 M * nkukard yea, in this case its just 2.6.32.25-vserver 1341354959 M * nkukard its been changed when building the kernel 1341355172 M * Bertl btw setreuid has a bunch of cases where it returns EPERM 1341355243 M * nkukard right, if I chroot, loging as the user and run the app, it works, but if I ssh into the vserver and run it I get the EPERM 1341355260 M * nkukard I'm trying to figure out if I can see from /proc/x/status what the app is running with in terms of caps 1341355972 M * Bertl assuming that no LSM is enabled, I'd look for some kind of dropped capabilitie and/or messed up (e/r)uids 1341356404 M * daniel_hozac note that setreuid will never work with strace 1341356417 M * daniel_hozac so you're probably chasing the wrong thing there. 1341356434 M * Bertl good point 1341356447 M * nkukard ah 1341356542 M * nkukard http://pastebin.com/zUdnJj9S 1341356558 M * nkukard 2 lines from the bottom I ran without strace 1341357098 M * Bertl check the app where the error message is generated and what might cause it 1341357201 Q * ensc|w Ping timeout: 480 seconds 1341357615 J * ensc|w ~ensc@www.sigma-chemnitz.de 1341359100 J * bergerx ~bergerx@static-94-33-32-60.clienti.tiscali.it