similar to: How can I get the pid mapping relationship between the container and the native OS

I'd also like to point out that, when setting up a linux container with _just_ the lxc tools available in ubuntu, all that was needed to be done was whitelist the appropriate cgroups for the Android devices. The lxc tools seem to create the appropriate character devices when you detach/attach the Android device via USB, and the Android `adb` tool recognized the devices. So, a tangential

Can you see the USB device in the lxc? My answering assumes you cannot see the USB device in the container. If you can see it, please ignore my answer. According to my understanding, “hostdev mode='subsystem' “ is used with KVM only. We use this if we want to passthrough a device to a KVM virtual machine. After doing that, the host machine will lose control of the device. If container is

So, if I remove the hostdev element related to the USB device, running `lsusb` and `usb-devices` inside the container lists the proper device, but I do not see the appropriate character device bus for the Android device under /dev/bus/usb. We are not entirely sure that `lsusb` and `usb-devices` queries the appropriate location - it somehow gets information from the host, not the guest container,

Dear all, I am not clear about the 'vcpu' element for CPU allocation. I allocated 1 vCPU to my container, after I started the container, I ran 4 computation-intensive tasks on the container. And I found all the 4 physical core are 100% used (my host has 4 physical cores and no other application ran on the host except the container). That is, all available cores were used by the container.

Dear Fil, I am not sure if my answer can help you. I had ever asked a similar question to Daniel and I was using a thirty-party card. As a container uses a shared kernel with the host, so hostdev mode='subsystem' doesn't make sense. Maybe you can try to use hostdev mode='capabilities'. Please see http://libvirt.org/formatdomain.html#elementsHostDevCaps Hope this helps Cheng

Hi, I'm not libvirt expect. My guess is that some vcpu settings only apply to KVM/qemu backend. LXC is quite different from them. If setting vcpu# is not effective for LXC container, you may need to use cgroups. -- Thanks, Yuanle On Mon, Mar 17, 2014 at 2:32 PM, WANG Cheng D < Cheng.d.Wang@alcatel-sbell.com.cn> wrote: > Dear all, > > I am not clear about the

Libvirt is in the standard repositories... Why don't you just install it with ex tasksel? Van: WANG Cheng D [mailto:Cheng.d.Wang@alcatel-sbell.com.cn] Verzonden: donderdag 26 maart 2015 16:06 Aan: libvirt-users@redhat.com Onderwerp: [libvirt-users] Installation of libvirt on Ubuntu Dear all, I wonder if libvirt can work on Ubuntu. As you know, it is very convenient to install libvirt on

Dear all, I wonder if libvirt can work on Ubuntu. As you know, it is very convenient to install libvirt on Fedora. I can download the libvirt packet from libvirt.org and install it using command ./configure, make and make install. As I am going to install an open source simulation system named openairinterface, Ubuntu is the recommended OS, and I want to install the simulation system in libvirt

Dear all, When I run my application in a container, sometimes the thread can exit correctly, sometime a "core dump" error occurs when it exits. But the application always exits correctly when it runs in a native fedora. Malloc() and free() functions are used in my application. I wonder if this is the reason for the core dump error. Is the thread exit event treated differently by the

Hi Chang Weng, Thanks for answering! Could you explain why subsystem=USB does not make sense? I was simply following the documentation for USB devices (right above the section you linked to). For what it's worth, I've also tried mounting the specific character device related to the appropriate android device into the container in the way you suggest, with no luck; `adb` still does not

On Mon, Dec 22, 2014 at 08:40:03AM +0000, WANG Cheng D wrote: >Dear all, > >I want my container to run on the third CPU core and I define this by the following xml scrpits: ><vcpu placement="static" cpuset="3">1</vcpu> > >When I run my application in a container terminal, I can see the application runs on the third core as expected. >When I run

Dear Daniel, You said "use a real init system like systemd to spawn off the things", do you mean a full OS distro container? I installed the container using "yum --installroot=/root/fedora19lxc --releasever=19 install -y openssh, bridge-utils" which is described at https://www.redhat.com/archives/libvirt-users/2013-August/msg00007.html. I don't know how to make systemd

Dear all, What I want to do is to share a hugepage memory segment between the host and a container (I am trying to use intel DPDK package in container). For a normal memory (4k page), the memory sharing can be achieved by memory-mapped I/O (mmap())method with the same disk file on the host exposed to the container (that is, the host and the container share the same disk file). my questions are:

Hey guys, When I destroy a lxc container, It reported an error with details as following: root@debian:~$ vir destroy 901eef81-06b2-4252-aeef-53b87c402585 error: Failed to destroy domain 901eef81-06b2-4252-aeef-53b87c402585 error: internal error Some container PIDs refused to die root@debian:~$ vir version Compiled against library: libvirt 1.1.0 Using library: libvirt 1.1.0 Using API: LXC

Dear Daniel, As you know, following your suggestion, I can expose shared memory to two containers hosted by the same machine, thus the two containers can exchange data. Now I have a new problem: how does the containers avoid concurrent accesses to the shared memory? In a native system, this can be achieved by using semaphore. But I don't know how the two containers can see the same semaphore.

Dear Daniel, Thank you for your email. According to your suggestion, I tried to use the shared memory filesystem for data exchange between the host machine and the container. The steps are as follows: 1) create an empty file in the host machine (/tmp1/file1). 2) run #chmod 777 /tmp1 and #chmod 777 /tmp1/file1. 3) expose /tmp1 to the container by configuring the container with the following XML

Dear All, I want to start my application automatically after I start the container with a virsh command " virsh -c lxc:/// start mycontainer" How can I achieve this? Thank you in advance. Cheng

Dear all, There are two containers hosted by my host machine. And the host machine has two same 1G NICs. I want to assign a dedicated NIC to one container. The host and the other container share the second NIC. How can I achieve this? Any comments will be highly appreciated. Cheng

Dear all, I installed 8 containers using "yum --installroot=/root/fedora16lxc" to create the container filesystem. Now I want to manage these containers with openstack. So I need to deploy these containers by openstack which needs the container image. Can anyone tell me how to export my existing container's image? I am not sure if openstack can create container without an image

Dear all, I want my container to run on the third CPU core and I define this by the following xml scrpits: <vcpu placement="static" cpuset="3">1</vcpu> When I run my application in a container terminal, I can see the application runs on the third core as expected. When I run my application using lxc-enter-namespace, the CPU pinning doesn't take effect, i.e.,