[Lxc-users] Dependencies Use Cases
I would like to use LXC on Ubuntu 11.10, but there are a few points that I'm having a hard time figuring out. I have been reading documentation, but it's a bit spread out and differs a bit from site to site. Please share your opinion and perspective about the dependencies and recommends of the LXC package in Ubuntu, particularly what you can imagine the use cases would be for installing or not installing recommended packages. From Aptitude on Ubuntu amd64 11.10 lxc - Depends (2) -- libc6 (= 2.8) -- libcap (= 2.10) - Recommends (3) -- cgroup-lite | cgroup-bin -- debootstrap -- libcap2-bin cgroup-bin - Depends (3) -- libc6 (= 2.7) -- libcgroup1 -- upstart-job cgroup-lite - Depends (1) -- upstart-job - Conflicts -- cgroup-bin From my perspective, having only read about LXC, it seems like libcap and libcgroup1 should be dependencies of LXC while libcap2-bin and cgroup-bin should be recommends of LXC. No offence to the creator of cgroup-lite, but from my limited perspective it seems like uninstalling LXC is a better solution than using cgroup-lite. Description of cgroup-lite. http://www.ubuntuupdates.org/packages/show/365681 Rational behind cgroup-lite. https://bugs.launchpad.net/ubuntu/+source/libcgroup/+bug/829628 -- RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 ___ Lxc-users mailing list Lxc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lxc-users
Re: [Lxc-users] Live Migration of LXC
On 10/24/2011 02:07 PM, Ulli Horlacher wrote: On Mon 2011-10-24 (12:03), Greg Kurz wrote: C/R and live migration is a complicated matter for LXC containers. I have assumed nothing else... Give BLCR a try : https://ftg.lbl.gov/CheckpointRestart/CheckpointDownloads.html It doesn't require a kernel patch, although it's clearly peeking and poking with the kernel internals. O.8.4 builds smoothly on debian, f14, and 2.6.38. I don't think it would be too complex to hook BLCR in C/R backend available in lxc. No status for the moment... I guess people who really want migration should participate Not every LXC (admin-)user is a kernel hacker, too. I am fluent in Perl programming, but not in C. at least to show kernel maintainers there's a demand for it. How can we do this? Send mass e-mails (spam) to the kernel maintainers? :-) That has been done already. Cheers, C. -- RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 ___ Lxc-users mailing list Lxc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lxc-users
Re: [Lxc-users] Dependencies Use Cases
On 11/04/2011 09:48 AM, Daniel Baumann wrote: in debian it's exactely like that (except that there's no relation to cgroup-bin yet, though i've added that in git to recommends for my next upload of lxc to debian). removed cgroup-bin from recommends again, cgroup-bin in debian is not ready and needs several updates first (remember: recommends means that it's by default installed on users systems). -- Address:Daniel Baumann, Donnerbuehlweg 3, CH-3012 Bern Email: daniel.baum...@progress-technologies.net Internet: http://people.progress-technologies.net/~daniel.baumann/ -- RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 ___ Lxc-users mailing list Lxc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lxc-users
Re: [Lxc-users] Dependencies Use Cases
On 11/04/2011 01:16 PM, Huang Liang wrote: Check out toft: https://github.com/exceedhl/toft. It provides rpm and deb packages which already handles the dependencies on centos and ubuntu. why would one want this instead of using lxc from your distributions repository? Moreover, it packages the bind and dhcp setup on the host machine and ships with pre-created images, saves a lot of time of hassling around these issues. that particular 'problem' we're going to solve in debian within about a week when lxc provides linux-container (a generic version of something similar what lxcguest in ubuntu and for ubuntu-only does) and live-build therefore can build proper system images for lxc containers that are shipped through .debs and which are going to be prefered over caches in /var/cache/lxc in debians lxc package. don't know what ubuntu has in mind for such use cases. -- Address:Daniel Baumann, Donnerbuehlweg 3, CH-3012 Bern Email: daniel.baum...@progress-technologies.net Internet: http://people.progress-technologies.net/~daniel.baumann/ -- RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 ___ Lxc-users mailing list Lxc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lxc-users
Re: [Lxc-users] Dependencies Use Cases
Quoting Daniel Baumann (daniel.baum...@progress-technologies.net): On 11/04/2011 01:16 PM, Huang Liang wrote: Check out toft: https://github.com/exceedhl/toft. It provides rpm and deb packages which already handles the dependencies on centos and ubuntu. why would one want this instead of using lxc from your distributions repository? Moreover, it packages the bind and dhcp setup on the host machine and ships with pre-created images, saves a lot of time of hassling around these issues. that particular 'problem' we're going to solve in debian within about a week when lxc provides linux-container (a generic version of something similar what lxcguest in ubuntu and for ubuntu-only does) and live-build therefore can build proper system images for lxc containers that are shipped through .debs and which are going to be prefered over caches in /var/cache/lxc in debians lxc package. don't know what ubuntu has in mind for such use cases. We aim to fix the two things that lxcguest is currently papering over so that the same unmodified ubuntu install can be used in kvm, lxc, or on bare metal. -serge -- RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 ___ Lxc-users mailing list Lxc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lxc-users
Re: [Lxc-users] /proc/process id/ns is not found
Thank you both Greg and Stephenie. (But still problem) I could build Linux 2.6.83.2 with lxc patches. I could make lxc-0.7.4 work with it. (lxc-0.7.5 did not work.) I mean it doesn't complain any more. However, I'm not sure if it really works. I did the follwoing: $ lxc-attach -n foo -- /bin/cat /etc/fstab But it shows host's /etc/fstab (not foo's /etc/fstab). I did a few other things. But the outputs are the host's (not the container's). Did I do something wrong? David. -- Dr. Dong-In David Kang Computer Scientist USC/ISI - Original Message - From: Greg Kurz gk...@fr.ibm.com To: Dong-In David Kang dk...@isi.edu Cc: lxc-users@lists.sourceforge.net Sent: Thursday, November 3, 2011 4:52:41 PM Subject: Re: [Lxc-users] /proc/process id/ns is not found On Thu, 2011-11-03 at 13:08 -0700, Dong-In David Kang wrote: Hello, I'm trying to use lxc-attach. The kernel I'm running is 2.6.38.8. I'm using lxc-0.7.5. When I run lxc-attach it tries to open /proc/process id/ns/..., but I cannot find it. Did I configure the Linux kernel in a wrong way? I could run an lxc instance. But I cannot use lxc-attach at the host. Any help? Hi, For lxc-attach to work, you need to patch your kernel with: http://lxc.sourceforge.net/patches/linux/2.6.38/2.6.38.2-lxc1/patches/ If you're running application containers (AKA. with lxc-execute(1)), you can give a try to the lxc patches discussed in this thread: https://sourceforge.net/mailarchive/message.php?msg_id=28283923 They introduce a userland only solution to run an extra commands in an already running container. Thanks, David. -- Dr. Dong-In David Kang Computer Scientist USC/ISI -- RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 ___ Lxc-users mailing list Lxc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lxc-users -- Gregory Kurz gk...@fr.ibm.com Software Engineer @ IBM/Meiosys http://www.ibm.com Tel +33 (0)534 638 479 Fax +33 (0)561 400 420 Anarchy is about taking complete responsibility for yourself. Alan Moore. -- RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 ___ Lxc-users mailing list Lxc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lxc-users
Re: [Lxc-users] mknod inside a container
On 11/04/2011 03:34 PM, Gordon Henderson wrote: I have a container that's used to build a Linux image for an embedded device - and as part of the build script, it creates /dev/ via a sequence of mknod commands Which all fail )-: There are no cap.drop lines in the contianers config files and I'm currently working round this by doing it on the host and copying the directory from the host to the container but I'd really rather do it inside the container... So what have I missed, or is it simply not possible? You probably have mknod restrictions through the lxc configuration file. Check for lxc.cgroup.devices.* in the configuration file and comment them all. Cheers -- Daniel -- RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 ___ Lxc-users mailing list Lxc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lxc-users
Re: [Lxc-users] /proc/process id/ns is not found
On 11/04/2011 03:37 PM, Dong-In David Kang wrote: I could build Linux 2.6.83.2 with lxc patches. I could make lxc-0.7.4 work with it. (lxc-0.7.5 did not work.) I mean it doesn't complain any more. However, I'm not sure if it really works. I did the follwoing: $ lxc-attach -n foo -- /bin/cat /etc/fstab But it shows host's /etc/fstab (not foo's /etc/fstab). I did a few other things. But the outputs are the host's (not the container's). what about : $ lxc-attach -n foo -- ps -ef --forest C. -- RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 ___ Lxc-users mailing list Lxc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lxc-users
Re: [Lxc-users] Dependencies Use Cases
Can LXC use cgroups without libcgroup? For that matter, just to be clear, can LXC use cgroups without cgroup-bin? In what use case would using LXC without cgroups make sense? Aren't cgroups fundamental to LXC? If cgroups are fundamental to LXC, then whatever is needed to make LXC capable of using cgroups needs to be a requirement regardless of whether that requirement is stable. I don't claim to know specifically which functionality and or packages are the real underlying dependencies, but it seems to me that it doesn't make sense to me to remove a real dependencies, whatever they may actually be, based on stability. If the dependency is unstable, then the dependent is unstable. Stability should not come at the price of making a package purposeless. I'm not saying that is actually what is happening, but based on my presumptions, which presumptions I am actively asking you to correct, that is what appears to have occurred in Ubuntu. Thank you everyone for your contributions. LXC is an awesome technology and I hope to have it up and running soon. serge, as a fellow member of the Ubuntu community, please do not refer to others' efforts as 'papering over' even if it perhaps is in response to someone saying that your efforts are pointless or stupid. I apologize for my wording about cgroup-lite. I don't understand the rational behind it based on my very likely incorrect presumptions, but I want to understand. Thank you for your contributions. Alex Eagar On 11/4/11, Serge Hallyn serge.hal...@canonical.com wrote: Quoting Daniel Baumann (daniel.baum...@progress-technologies.net): On 11/04/2011 01:16 PM, Huang Liang wrote: Check out toft: https://github.com/exceedhl/toft. It provides rpm and deb packages which already handles the dependencies on centos and ubuntu. why would one want this instead of using lxc from your distributions repository? Moreover, it packages the bind and dhcp setup on the host machine and ships with pre-created images, saves a lot of time of hassling around these issues. that particular 'problem' we're going to solve in debian within about a week when lxc provides linux-container (a generic version of something similar what lxcguest in ubuntu and for ubuntu-only does) and live-build therefore can build proper system images for lxc containers that are shipped through .debs and which are going to be prefered over caches in /var/cache/lxc in debians lxc package. don't know what ubuntu has in mind for such use cases. We aim to fix the two things that lxcguest is currently papering over so that the same unmodified ubuntu install can be used in kvm, lxc, or on bare metal. -serge -- RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 ___ Lxc-users mailing list Lxc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lxc-users -- RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 ___ Lxc-users mailing list Lxc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lxc-users
Re: [Lxc-users] /proc/process id/ns is not found
Never mind. I put the wrong PID. I got PID of the lxc process. But I had to use init process of the lxc process, which is usually 1 + PID of the lxc process. Thanks, David. -- Dr. Dong-In David Kang Computer Scientist USC/ISI - Original Message - From: Dong-In David Kang dk...@isi.edu To: Cedric Le Goater legoa...@free.fr Cc: lxc-users@lists.sourceforge.net Sent: Friday, November 4, 2011 11:15:05 AM Subject: Re: [Lxc-users] /proc/process id/ns is not found Here is the output of $ lxc-attach -n foo -- ps -ef --forest. (I've changed lxc-attach a little bit so that I can provide processor id instead of the name of lxc process. I've started the lxc process using libvirt. The process id of the lxc process is 4921.) [root@gpu2 lxc]# lxc-attach -n 4921 -- ps -ef --forest lxc-attach: No such file or directory - failed to exec 'ps' [root@gpu2 lxc]# lxc-attach -n 4921 -- /bin/ps -ef --forest UIDPID PPID C STIME TTY TIME CMD root 2 0 0 Nov03 ?00:00:00 [kthreadd] root 3 2 0 Nov03 ?00:00:00 \_ [ksoftirqd/0] root 6 2 0 Nov03 ?00:00:00 \_ [migration/0] root 7 2 0 Nov03 ?00:00:00 \_ [watchdog/0] root 8 2 0 Nov03 ?00:00:00 \_ [migration/1] root 9 2 0 Nov03 ?00:00:00 \_ [kworker/1:0] root10 2 0 Nov03 ?00:00:00 \_ [ksoftirqd/1] root12 2 0 Nov03 ?00:00:00 \_ [watchdog/1] root13 2 0 Nov03 ?00:00:00 \_ [migration/2] root14 2 0 Nov03 ?00:00:00 \_ [kworker/2:0] root15 2 0 Nov03 ?00:00:00 \_ [ksoftirqd/2] root16 2 0 Nov03 ?00:00:00 \_ [watchdog/2] root17 2 0 Nov03 ?00:00:00 \_ [migration/3] root18 2 0 Nov03 ?00:00:00 \_ [kworker/3:0] root19 2 0 Nov03 ?00:00:00 \_ [ksoftirqd/3] root20 2 0 Nov03 ?00:00:00 \_ [watchdog/3] root21 2 0 Nov03 ?00:00:00 \_ [migration/4] root23 2 0 Nov03 ?00:00:00 \_ [ksoftirqd/4] root24 2 0 Nov03 ?00:00:00 \_ [watchdog/4] root25 2 0 Nov03 ?00:00:00 \_ [migration/5] root26 2 0 Nov03 ?00:00:01 \_ [kworker/5:0] root27 2 0 Nov03 ?00:00:00 \_ [ksoftirqd/5] root28 2 0 Nov03 ?00:00:00 \_ [watchdog/5] root29 2 0 Nov03 ?00:00:00 \_ [migration/6] root30 2 0 Nov03 ?00:00:00 \_ [kworker/6:0] root31 2 0 Nov03 ?00:00:00 \_ [ksoftirqd/6] root32 2 0 Nov03 ?00:00:00 \_ [watchdog/6] root33 2 0 Nov03 ?00:00:00 \_ [migration/7] root34 2 0 Nov03 ?00:00:00 \_ [kworker/7:0] root35 2 0 Nov03 ?00:00:00 \_ [ksoftirqd/7] root36 2 0 Nov03 ?00:00:00 \_ [watchdog/7] root37 2 0 Nov03 ?00:00:00 \_ [cpuset] root38 2 0 Nov03 ?00:00:00 \_ [khelper] root39 2 0 Nov03 ?00:00:00 \_ [netns] root40 2 0 Nov03 ?00:00:00 \_ [sync_supers] root41 2 0 Nov03 ?00:00:00 \_ [bdi-default] root42 2 0 Nov03 ?00:00:00 \_ [kintegrityd] root43 2 0 Nov03 ?00:00:00 \_ [kblockd] root44 2 0 Nov03 ?00:00:00 \_ [ata_sff] root45 2 0 Nov03 ?00:00:00 \_ [khubd] root46 2 0 Nov03 ?00:00:00 \_ [md] root47 2 0 Nov03 ?00:00:00 \_ [kworker/1:1] root48 2 0 Nov03 ?00:00:01 \_ [kworker/2:1] root49 2 0 Nov03 ?00:00:00 \_ [kworker/3:1] root50 2 0 Nov03 ?00:00:02 \_ [kworker/4:1] root52 2 0 Nov03 ?00:00:01 \_ [kworker/6:1] root53 2 0 Nov03 ?00:00:00 \_ [kworker/7:1] root54 2 0 Nov03 ?00:00:00 \_ [khungtaskd] root55 2 0 Nov03 ?00:00:02 \_ [kswapd0] root56 2 0 Nov03 ?00:00:00 \_ [ksmd] root57 2 0 Nov03 ?00:00:00 \_ [khugepaged] root58 2 0 Nov03 ?00:00:00 \_ [fsnotify_mark] root59 2 0 Nov03 ?00:00:00 \_ [aio] root60 2 0 Nov03 ?00:00:00 \_ [crypto] root65 2 0 Nov03 ?00:00:00 \_ [kthrotld] root67 2 0 Nov03 ?00:00:00 \_ [kpsmoused] root89 2 0 Nov03 ?00:00:03 \_ [kworker/0:2] root 235 2 0 Nov03 ?00:00:00 \_ [scsi_eh_0] root 236 2 0 Nov03 ?00:00:00 \_ [scsi_eh_1] root 241 2 0 Nov03 ?00:00:00 \_ [scsi_eh_2] root 242 2 0 Nov03 ?00:00:00 \_ [scsi_eh_3] root 243 2 0 Nov03 ?00:00:00 \_ [scsi_eh_4] root
Re: [Lxc-users] mknod inside a container
On Fri, 4 Nov 2011, Daniel Lezcano wrote: On 11/04/2011 03:34 PM, Gordon Henderson wrote: I have a container that's used to build a Linux image for an embedded device - and as part of the build script, it creates /dev/ via a sequence of mknod commands Which all fail )-: There are no cap.drop lines in the contianers config files and I'm currently working round this by doing it on the host and copying the directory from the host to the container but I'd really rather do it inside the container... So what have I missed, or is it simply not possible? You probably have mknod restrictions through the lxc configuration file. Check for lxc.cgroup.devices.* in the configuration file and comment them all. Yup. That was it, thanks! I had it in my mind that it was capabilities rather than simple devices stuff. Cheers, Gordon -- RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 ___ Lxc-users mailing list Lxc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lxc-users
Re: [Lxc-users] Dependencies Use Cases
Quoting Alex Eagar (alexea...@gmail.com): Can LXC use cgroups without libcgroup? For that matter, just to be clear, can LXC use cgroups without cgroup-bin? LXC doesn't need anything from cgroup-bin, and, if it did, cgroup-bin could not deliver. (see below) In what use case would using LXC without cgroups make sense? Aren't cgroups fundamental to I think you misunderstand cgroup-bin. The point of cgroup-bin is to try and catch applications/daemons as they start and classify them into cgroups according to a configuration. However because tasks are classified by placing their pids one at a time into a file, there are cases where it misses tasks, and it's not entirely reliable. LXC controls cgroups (the kernel feature) itself through the cgroup filesystem. cgroup-bin is not needed for this. The cgroups just need to be composed in a (set of) cgroup mount(s) somewhere. happening, but based on my presumptions, which presumptions I am actively asking you to correct, that is what appears to have occurred in Ubuntu. Hopefully the above explained why that's not what happened. serge, as a fellow member of the Ubuntu community, please do not refer to others' efforts as 'papering over' even if it perhaps is in I wrote lxcguest. 'Papering over' is not meant as a put-down. The point is that there are things in a stock Ubuntu install which stop a container from booting. For each of those, the right thing to do is to update the packages involved so that they can work just as well in a container as on hardware/kvm. But for a first step, I chose to create a package to hide the problems. In part, that gave us a better chance to figure out what the real problems were. Currently there are (if I'm thinking right), at core, two: 1. the need for the lxc-monitor to watch /run/utmp in the container to detect reboot/shutdown. That means the guest can't mount tmpfs on /run, which suddenly creates a whole set of issues. Daniel is hoping to resend a kernel patchset this week or next which well let us not do that. 2. mountall needs to not mount certain things in a container at boot. Here is where I almost literally paper over :) : lxcguest just bind-mounts a different file over /lib/init/fstab to make mountall do what we want. This can break upgrades, when they want to overwrite /lib/init/fstab. So I intend to fix mountall so we don't need that. -serge -- RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 ___ Lxc-users mailing list Lxc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lxc-users
Re: [Lxc-users] /proc/process id/ns is not found
On Fri, 2011-11-04 at 08:15 -0700, Dong-In David Kang wrote: Here is the output of $ lxc-attach -n foo -- ps -ef --forest. (I've changed lxc-attach a little bit so that I can provide processor id instead of the name of lxc process. I've started the lxc process using libvirt. The process id of the lxc process is 4921.) I'm not familiar with libvirt, but from the ps output below, it appears that the pid you're interested in is the one running /sbin/init (4922). It's the first process, with pid 1, in the container. The libvirt_lxc (4921) process seems to be the container parent. [root@gpu2 lxc]# lxc-attach -n 4921 -- ps -ef --forest lxc-attach: No such file or directory - failed to exec 'ps' [root@gpu2 lxc]# lxc-attach -n 4921 -- /bin/ps -ef --forest UIDPID PPID C STIME TTY TIME CMD root 2 0 0 Nov03 ?00:00:00 [kthreadd] root 3 2 0 Nov03 ?00:00:00 \_ [ksoftirqd/0] root 6 2 0 Nov03 ?00:00:00 \_ [migration/0] root 7 2 0 Nov03 ?00:00:00 \_ [watchdog/0] root 8 2 0 Nov03 ?00:00:00 \_ [migration/1] root 9 2 0 Nov03 ?00:00:00 \_ [kworker/1:0] root10 2 0 Nov03 ?00:00:00 \_ [ksoftirqd/1] root12 2 0 Nov03 ?00:00:00 \_ [watchdog/1] root13 2 0 Nov03 ?00:00:00 \_ [migration/2] root14 2 0 Nov03 ?00:00:00 \_ [kworker/2:0] root15 2 0 Nov03 ?00:00:00 \_ [ksoftirqd/2] root16 2 0 Nov03 ?00:00:00 \_ [watchdog/2] root17 2 0 Nov03 ?00:00:00 \_ [migration/3] root18 2 0 Nov03 ?00:00:00 \_ [kworker/3:0] root19 2 0 Nov03 ?00:00:00 \_ [ksoftirqd/3] root20 2 0 Nov03 ?00:00:00 \_ [watchdog/3] root21 2 0 Nov03 ?00:00:00 \_ [migration/4] root23 2 0 Nov03 ?00:00:00 \_ [ksoftirqd/4] root24 2 0 Nov03 ?00:00:00 \_ [watchdog/4] root25 2 0 Nov03 ?00:00:00 \_ [migration/5] root26 2 0 Nov03 ?00:00:01 \_ [kworker/5:0] root27 2 0 Nov03 ?00:00:00 \_ [ksoftirqd/5] root28 2 0 Nov03 ?00:00:00 \_ [watchdog/5] root29 2 0 Nov03 ?00:00:00 \_ [migration/6] root30 2 0 Nov03 ?00:00:00 \_ [kworker/6:0] root31 2 0 Nov03 ?00:00:00 \_ [ksoftirqd/6] root32 2 0 Nov03 ?00:00:00 \_ [watchdog/6] root33 2 0 Nov03 ?00:00:00 \_ [migration/7] root34 2 0 Nov03 ?00:00:00 \_ [kworker/7:0] root35 2 0 Nov03 ?00:00:00 \_ [ksoftirqd/7] root36 2 0 Nov03 ?00:00:00 \_ [watchdog/7] root37 2 0 Nov03 ?00:00:00 \_ [cpuset] root38 2 0 Nov03 ?00:00:00 \_ [khelper] root39 2 0 Nov03 ?00:00:00 \_ [netns] root40 2 0 Nov03 ?00:00:00 \_ [sync_supers] root41 2 0 Nov03 ?00:00:00 \_ [bdi-default] root42 2 0 Nov03 ?00:00:00 \_ [kintegrityd] root43 2 0 Nov03 ?00:00:00 \_ [kblockd] root44 2 0 Nov03 ?00:00:00 \_ [ata_sff] root45 2 0 Nov03 ?00:00:00 \_ [khubd] root46 2 0 Nov03 ?00:00:00 \_ [md] root47 2 0 Nov03 ?00:00:00 \_ [kworker/1:1] root48 2 0 Nov03 ?00:00:01 \_ [kworker/2:1] root49 2 0 Nov03 ?00:00:00 \_ [kworker/3:1] root50 2 0 Nov03 ?00:00:02 \_ [kworker/4:1] root52 2 0 Nov03 ?00:00:01 \_ [kworker/6:1] root53 2 0 Nov03 ?00:00:00 \_ [kworker/7:1] root54 2 0 Nov03 ?00:00:00 \_ [khungtaskd] root55 2 0 Nov03 ?00:00:02 \_ [kswapd0] root56 2 0 Nov03 ?00:00:00 \_ [ksmd] root57 2 0 Nov03 ?00:00:00 \_ [khugepaged] root58 2 0 Nov03 ?00:00:00 \_ [fsnotify_mark] root59 2 0 Nov03 ?00:00:00 \_ [aio] root60 2 0 Nov03 ?00:00:00 \_ [crypto] root65 2 0 Nov03 ?00:00:00 \_ [kthrotld] root67 2 0 Nov03 ?00:00:00 \_ [kpsmoused] root89 2 0 Nov03 ?00:00:03 \_ [kworker/0:2] root 235 2 0 Nov03 ?00:00:00 \_ [scsi_eh_0] root 236 2 0 Nov03 ?00:00:00 \_ [scsi_eh_1] root 241 2 0 Nov03 ?00:00:00 \_ [scsi_eh_2] root 242 2 0 Nov03 ?00:00:00 \_ [scsi_eh_3] root 243 2 0 Nov03 ?00:00:00 \_ [scsi_eh_4] root 244 2 0 Nov03 ?00:00:00 \_ [scsi_eh_5] root 245 2 0 Nov03 ?
[Lxc-users] mknod after instance creation?
Hi, Is it possible to do mknod after creation of an LXC instance? I need to do mknod not only at bootup time, but also at run-time. This is needed when I want to dynamically add devices to LXC instance. Is it possible? If it is, how can I do it? I've seen the case of mknod at bootup time of an LXC instance. But, I haven't seen the usage of mknod at run-time after boot-up. Is it the limitation of LXC? Thanks, David. -- Dr. Dong-In David Kang Computer Scientist USC/ISI -- RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 ___ Lxc-users mailing list Lxc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lxc-users
Re: [Lxc-users] mknod after instance creation?
On 11/05/2011 12:06 AM, Dong-In David Kang wrote: Hi, Is it possible to do mknod after creation of an LXC instance? I need to do mknod not only at bootup time, but also at run-time. This is needed when I want to dynamically add devices to LXC instance. Is it possible? If it is, how can I do it? I've seen the case of mknod at bootup time of an LXC instance. But, I haven't seen the usage of mknod at run-time after boot-up. Is it the limitation of LXC? Just comment out the lxc.cgroup.devices.* lines in the configuration file. -- RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 ___ Lxc-users mailing list Lxc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lxc-users
Re: [Lxc-users] Dependencies Use Cases
It would be great if lxc itself can solve these nuance problems, because I found at least more than one version of tutorial/howto to guide you run different containers on different host machines; and the currently shipped lxc distribution has not been tested in all container/host combinations. The advantage of shipping container images rather than building cache rootfs in templates are two - it is faster to extract a tar ball; and, as far as I know, it make it possible to run debian/ubuntu container on centos since centos does not have debootstrap and you can not create debian image on centos. Toft is merely a simple wrapper to help people test infrastructure code. It could be more simpler in the future if it does not have do things that should be in lxc. On Fri, Nov 4, 2011 at 11:27 PM, Daniel Baumann daniel.baum...@progress-technologies.net wrote: On 11/04/2011 01:16 PM, Huang Liang wrote: Check out toft: https://github.com/exceedhl/toft. It provides rpm and deb packages which already handles the dependencies on centos and ubuntu. why would one want this instead of using lxc from your distributions repository? Moreover, it packages the bind and dhcp setup on the host machine and ships with pre-created images, saves a lot of time of hassling around these issues. that particular 'problem' we're going to solve in debian within about a week when lxc provides linux-container (a generic version of something similar what lxcguest in ubuntu and for ubuntu-only does) and live-build therefore can build proper system images for lxc containers that are shipped through .debs and which are going to be prefered over caches in /var/cache/lxc in debians lxc package. don't know what ubuntu has in mind for such use cases. -- Address: Daniel Baumann, Donnerbuehlweg 3, CH-3012 Bern Email: daniel.baum...@progress-technologies.net Internet: http://people.progress-technologies.net/~daniel.baumann/ -- RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 ___ Lxc-users mailing list Lxc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lxc-users -- Yours Sincerely, 黄亮 --- Huang Liang, ThoughtWorks, Inc Tel: (+86)13911786684 Email: lhu...@thoughtworks.com -- RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 ___ Lxc-users mailing list Lxc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lxc-users