Users may want to run the init command of a container as a special
user / group. This is achieved by adding <inituser> and <initgroup>
elements. Note that the user can either provide a name or an ID to
specify the user / group to be used.
This commit also fixes a side effect of being able to run the command
as a non-root user: the user needs rights on the tty to allow shell
job control.
Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
Some containers may want the application to run in a special directory.
Add <initdir> element in the domain configuration to handle this case
and use it in the lxc driver.
Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
When running an application container, setting environment variables
could be important.
The newly introduced <initenv> tag in domain configuration will allow
setting environment variables to the init program.
Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
This is place as a sub-element of <source>, where other aspects of the
host-side connection to the network device are located (network or
bridge name, udp listen port, etc). It's a bit odd that the interface
we're configuring with this info is itself named in <target dev='x'/>,
but that ship sailed long ago:
<interface type='ethernet'>
<mac address='00:16:3e:0f:ef:8a'/>
<source>
<ip address='192.168.122.12' family='ipv4'
prefix='24' peer='192.168.122.1'/>
<ip address='192.168.122.13' family='ipv4' prefix='24'/>
<route family='ipv4' address='0.0.0.0'
gateway='192.168.122.1'/>
<route family='ipv4' address='192.168.124.0' prefix='24'
gateway='192.168.124.1'/>
</source>
</interface>
In practice, this will likely only be useful for type='ethernet', so
its presence in any other type of interface is currently forbidden in
the generic device Validate function (but it's been put into the
general population of virDomainNetDef rather than the
ethernet-specific union member so that 1) we can more easily add the
capability to other types if needed, and 2) we can retain the info
when set to an invalid interface type all the way through to
validation and report a proper error, rather than just ignoring it
(which is currently what happens for many other type-specific
settings).
(NB: The already-existing configuration of IP info for the guest-side
of interfaces is in subelements directly under <interface>, and the
name of the guest-side interface (when configurable) is in <guest
dev='x'/>).
(This patch had been pushed earlier in
commit fe6a77898a, but was reverted in
commit d658456530 because it had been
accidentally pushed during the freeze for release 2.0.0)
This is place as a sub-element of <source>, where other aspects of the
host-side connection to the network device are located (network or
bridge name, udp listen port, etc). It's a bit odd that the interface
we're configuring with this info is itself named in <target dev='x'/>,
but that ship sailed long ago:
<interface type='ethernet'>
<mac address='00:16:3e:0f:ef:8a'/>
<source>
<ip address='192.168.122.12' family='ipv4'
prefix='24' peer='192.168.122.1'/>
<ip address='192.168.122.13' family='ipv4' prefix='24'/>
<route family='ipv4' address='0.0.0.0'
gateway='192.168.122.1'/>
<route family='ipv4' address='192.168.124.0' prefix='24'
gateway='192.168.124.1'/>
</source>
</interface>
In practice, this will likely only be useful for type='ethernet', so
its presence in any other type of interface is currently forbidden in
the generic device Validate function (but it's been put into the
general population of virDomainNetDef rather than the
ethernet-specific union member so that 1) we can more easily add the
capability to other types, and 2) we can retain the info when set to
an invalid interface type all the way through to validation and report
a proper error, rather than just ignoring it (which is currently what
happens for many other type-specific settings).
(NB: The already-existing configuration of IP info for the guest-side
of interfaces is in subelements directly under <interface>, and the
name of the guest-side interface (when configurable) is in <guest
dev='x'/>).
This is identical to type='bridge', but without the "connect to a
bridge" part, so it can be handled by using the same functions (and
often even the same cases in switch statements), after renaming
virLXCProcessSetupInterfaceBridged() to virLXCProcessInterfaceTap()
and enhancing it to skip bridge-related items when brname == NULL.
To be truly useful, we need to support setting the ip address on the
host side veth as well as guest side veth (already supported for
type='bridge'), as well as setting the peer address for both.
The <script> element (supported by type='ethernet' in qemu) isn't
supported in this patch. An error is logged at domain start time if it
is encountered. This may be changed in a later patch.
This patch adds feature for lxc containers to inherit namespaces.
This is very similar to what lxc-tools or docker provides. Look
for "man lxc-start" and you will find that you can pass command
args as [ --share-[net|ipc|uts] name|pid ]. Or check out docker
networking option in which you can give --net=container:NAME_or_ID
as an option for sharing +namespace.
>From this patch you can add extra libvirt option to share
namespace in following way.
<lxc:namespace>
<lxc:sharenet type='netns' value='red'/>
<lxc:shareipc type='pid' value='12345'/>
<lxc:shareuts type='name' value='container1'/>
</lxc:namespace>
The netns option is specific to sharenet. It can be used to
inherit from existing network namespace.
Co-authored: Daniel P. Berrange <berrange@redhat.com>
Add the possibility to have more than one IP address configured for a
domain network interface. IP addresses can also have a prefix to define
the corresponding netmask.
If all features are set to default (including the capabilities policy),
but some capabilities are toggled, we need to output the <features>
element when formatting the config.
LXC network devices can now be assigned a custom NIC device name on the
container side. For example, this is configured with:
<interface type='network'>
<source network='default'/>
<guest dev="eth1"/>
</interface>
In this example the network card will appear as eth1 in the guest.
When idmap was added to LXC, we forgot to cover it in the testsuite.
The schema was missing an <element> layer, and as a result,
virt-xml-validate was failing on valid dumpxml output.
Reported by Eduard - Gabriel Munteanu on IRC.
* docs/schemas/domaincommon.rng (idmap): Include <idmap> element,
and support interleaves.
* tests/lxcxml2xmldata/lxc-idmap.xml: New file.
* tests/lxcxml2xmltest.c (mymain): Test it.
Signed-off-by: Eric Blake <eblake@redhat.com>
Since 76b644c when the support for RAM filesystems was introduced,
libvirt accepted the following XML:
<source usage='1024' unit='KiB'/>
This was parsed correctly and internally stored in bytes, but it
was formatted as (with an extra 's'):
<source usage='1024' units='KiB'/>
When read again, this was treated as if the units were missing,
meaning libvirt was unable to parse its own XML correctly.
The usage attribute was documented as being in KiB, but it was not
scaled if the unit was missing. Transient domains still worked,
because this was balanced by an extra 'k' in the mount options.
This patch:
Changes the parser to use 'units' instead of 'unit', as the latter
was never documented (fixing persistent domains) and some programs
(libvirt-glib, libvirt-sandbox) already parse the 'units' attribute.
Removes the extra 'k' from the tmpfs mount options, which is needed
because now we parse our own XML correctly.
Changes the default input unit to KiB to match documentation, fixing:
https://bugzilla.redhat.com/show_bug.cgi?id=1015689
The <filesystem> element can now accept a <driver type='nbd'/>
as an alternative to 'loop'. The benefit of NBD is support
for non-raw disk image formats.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Extend the <driver> element in filesystem devices to
allow a storage format to be set. The new attribute
uses 'format' to reflect the storage format. This is
different from the <driver> element in disk devices
which use 'type' to reflect the storage format. This
is because the 'type' attribute on filesystem devices
is already used for the driver backend, for which the
disk devices use the 'name' attribute. Arggggh.
Anyway for disks we have
<driver name="qemu" type="raw"/>
And for filesystems this change means we now have
<driver type="loop" format="raw"/>
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
This allows a container-type domain to have exclusive access to one of
the host's NICs.
Wire <hostdev caps=net> with the lxc_controller - when moving the newly
created veth devices into a new namespace, also look for any hostdev
devices that should be moved. Note: once the container domain has been
destroyed, there is no code that moves the interfaces back to the
original namespace. This does happen, though, probably due to default
cleanup on namespace destruction.
Signed-off-by: Bogdan Purcareata <bogdan.purcareata@freescale.com>
This patch removes the emulatorRequired field and associated
infrastructure from the virCaps object. Instead the driver specific
callbacks are used as this field isn't enforced by all drivers.
This patch implements the appropriate callbacks in the qemu and lxc
driver and moves to check to that location.
The <hostdev> device type has long had a redundant "mode"
attribute, which has always been "subsys". This finally
introduces a new mode "capabilities", which will be used
by the LXC driver for device assignment. Since container
based virtualization uses a single kernel, the idea of
assigning physical PCI devices doesn't make sense. It is
still reasonable to assign USB devices, but for assigning
arbitrary nodes in /dev, the new 'capabilities' mode is
to be used.
The first capability support is 'storage', which is for
assignment of block devices. Functionally this is really
pretty similar to the <disk> support. The only difference
is the device node name is identical in both host and
container namespaces.
<hostdev mode='capabilities' type='storage'>
<source>
<block>/dev/sdf1</block>
</source>
</hostdev>
The second capability support is 'misc', which is for
assignment of character devices. There is no existing
parallel to this. Again the device node is the same
inside & outside the container.
<hostdev mode='capabilities' type='misc'>
<source>
<char>/dev/input/event3</char>
</source>
</hostdev>
The reason for keeping the char & storage devices
separate in the domain XML, is to mirror the split
in the node device XML. NB the node device XML does
not yet report character devices, but that's another
new patch to come
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
<vcpu> is not an optional node. The value for its 'placement'
actually always defaults to 'static' in the underlying codes.
(Even no 'cpuset' and 'placement' is specified, the domain
process will be pinned to all the available pCPUs).
Pass argv to the init binary of LXC, using a new <initarg> element.
* docs/formatdomain.html.in: Document <os> usage for containers
* docs/schemas/domaincommon.rng: Add <initarg> element
* src/conf/domain_conf.c, src/conf/domain_conf.h: parsing and
formatting of <initarg>
* src/lxc/lxc_container.c: Setup LXC argv
* tests/Makefile.am, tests/lxcxml2xmldata/lxc-systemd.xml,
tests/lxcxml2xmltest.c, tests/testutilslxc.c,
tests/testutilslxc.h: Test parsing/formatting of LXC related
XML parts
