Changes

Jump to: navigation, search

Vzctl for upstream kernel

2,101 bytes added, 14:23, 22 June 2015
rename article
Since version 4.0, {{DISPLAYTITLE: vzctl tool can be used with for upstream (non-OpenVZ) Linux kernels (that essentially means any recent 3.x kernel). At the moment, it provides just basic functionality.It is currently possible to create and start a container with the same steps as one would use for a normal OpenVZ container. Other features may be present with limited functionality, while some are not present at all.}}
{{Warning|'''This article describes using OpenVZ tool vzctl as an alternative to LXC tools.''' Recent vzctl releases (starting from version 4.0) can be used with upstream (non-OpenVZ) Linux kernels (that essentially means any recent 3.x kernel). At the moment, it provides just basic functionality.It is currently possible to create, start and stop a container with the same steps as one would use for a normal OpenVZ container. Other features may be present with limited functionality, while some are not present at all. We appreciate all bug reports, please file to [http://bugzilla.openvz.org/enter_bug.cgi?component=vzctl bugzilla]. Running vzctl on upstream kernels is considered an experimental feature. See [[#LimitatonsLimitations]] below.}}
== Installation ==
=== Networking ===
{{Note|IP mode networking (--ipadd / --ipdel) is currently not supported}}
 
Networking is available through the switches <code>--netdev_add</code>, <code>--netif_add</code>, and their respective deletion counterparts.
Unfortunately now it requires some manual configuration.
 
== Bridged networking ==
 
The following example assumes
* you already have a bridge configured on the host system
* bridge interface name is virbr0
* CT is running Red Hat like distro (CentOS)
 
vzctl set $CTID --netif_add eth0,,,,virbr0 --save
 
echo "NETWORKING=yes" > /vz/private/$CTID/etc/sysconfig/network
Networking is available through the switches cat << EOF > /vz/private/$CTID/etc/sysconfig/network-scripts/ifcfg-netdev_add, --netif_add, and their respective deletion counterparts.eth0 DEVICE=eth0 BOOTPROTO=dhcp ONBOOT=yesIn this case [[Virtual Ethernet device]] is added in [[CT]]. EOF
One life hack may be useful while "vzctl enter" doesn't work. If you use DHCP and don't know which an IP address is in CT, you can find it in /vz/root/[start $CTID]/var/log/message.
After this, you can find CT IP mode networking (--ipadd / --ipdel) is currently not supported.using this: # ip netns exec $CTID ip address list
== Limitations ==
{{Note|We recommend using [[Download/kernel/rhel6|OpenVZ kernel]] for features, stability and security}} The following vzctl commands are not working at allwith the non-OpenVZ kernel
* <code>quotaon</code>/<code>quotaoff</code>/<code>quotainit</code> (vzquota-specific)
* <code>convert</code>, <code>compact</code>, <code>snapshot*</code> (ploop-specific)
* <code>console</code> (needs a virtual /dev/console, /dev/ttyN device)
* <code>enter</code>, <code>exec</code> and <code>runscript</code> (need pidns entering support)* <code>chkpnt</code>, <code>restore</code> (currently need OpenVZ-kernel-specific checkpointing, [http://cruicriu.org/ CRIU] will be supported later) The following commands have severe limitations:* <code>stop</code>. A container can be stopped from inside (say if one is connected to CT over ssh) in case the underlying kernel supports rebooting a PID namespace (> 3.4). Using vzctl, the "stop" command is not supported, unless accompanied by the --fast switch, which will simply forceably kill all processes in the container.
The following binaries are not ported to work on top of upstream kernel:
=== Resource management ===
Setting With non-OpenVZ kernel, setting resources like <code>--ram</code> and <code>--cpuunits</code> workworks, but there their effect is dependent on what the current kernel supports, through the cgroups subsystem. When a particular cgroup file is present, it will be used. Currently, vzctl will search for the following files:
* cpu.cfs_quota_us
* cpu.shares
* memory.kmem.tcp.limit_in_bytes
== Building ==
== Building ==In case you don't want to use packages provided by OpenVZ (available from [[Download/vzctl]]), but rather would like to compile vzctl from sources, read on.
=== Dependencies ===
If you are living on the bleeding edge, get vzctl sources from git. Then run autogen.sh to recreate auto* files:
git clone git<nowiki>https://gitsrc.openvz.org/pubscm/ovzl/vzctl.git</nowiki>
cd vzctl
./autogen.sh
=== Compile ===
Usual <code>./conficonfigure && make</code> should do. But you probably want to specify more options. It makes sense to: * enable cgroup support* add <code>--without-ploop</code> (unless you want [[ploop]] compiled it) because otherwise you will need ploop lib headers (available from [[Download/ploop]]).* enable bash completion support* set prefix to /usr See <code>./configure --help</code> output for more details and options available. So, the command will look like:  $ ./configure --with-cgroup --without-ploop --enable-bashcomp --prefix=/usr $ make -j4 === Install ===  # make install vzctl is also bundled in some Linux distributions, so you can install vzctl using native distro tools (i.e. your package manager): * vzctl in [https://launchpad.net/ubuntu/+source/vzctl Ubuntu Linux] [https://help.ubuntu.com/community/OpenVZ Description]* vzctl-core in [https://apps.fedoraproject.org/packages/vzctl-core Fedora Linux] [https://fedoraproject.org/wiki/QA:Testcase_vzctl_base Description]* sys-cluster/vzctl in [http://packages.gentoo.org/package/sys-cluster/vzctl Gentoo Linux]* vzctl in [https://packages.debian.org/search?keywords=vzctl Debian Linux]  == Known issues and workarounds == === A container doesn't boot and udevd is in a process list ===udev doesn't work, because <code>uevent</code>s are not virtualized yet. If you don't know how to disable it, you can remove the udev package. === <code>vzctl enter</code> doesn't work === You see this when trying to use <code>vzctl enter</code>:  Unable to open pty: No such file or directory If a CT is executed in a user namespace, devpts must be mounted with the newinstance option. You can add this option in container's <code>/etc/fstab</code> file.
t makes sense to add <code>--without-ploop</code> (unless you want ploop compiled it) because otherwise you will need ploop lib headers.== See also ==
$ ./configure --* [[OpenVZ with-cgroup --without-ploopupstream kernel]]

Navigation menu