Place an ephemeral container started with -e flag on a tmpfs. Restrictions are
that you cannot request the data to be kept while placing the container on a
tmpfs, that either overlay or aufs backing storage must be used, and that the
storage backend of the original container must be a directory.
For ephemeral snapshots backed by overlay or aufs filesystems, a fresh tmpfs
is mounted over the containers directory if the user requests it. This should
be the easiest options. Anything else would require us to change the current
mount-layout of overlay and aufs snapshots. (A standard overlay or aufs
snapshot clone currently has the layout:
/var/lib/lxc/CLONE_SNAPSHOT/delta0 <-- upperdir
/var/lib/lxc/CLONE_SNAPSHOT/rootfs
/var/lib/lxc/CLONE_SNAPSHOT/olwork
/var/lib/lxc/CLONE_SNAPSHOT/olwork/work <-- workdir
with the lowerdir being
/var/lib/lxc/CLONE_PARENT/rootfs
The fact that upperdir and workdir are not placed in a common subfolder under
the container directory has the consequence that we cannot simply mount a fresh
tmpfs under upperdir and workdir because overlay expects them to be on the same
filesystem.)
Because we mount a fresh tmpfs over the directory of the container the updated
/etc/hostname file created during the clone residing in the upperdir (currently
named "delta0" by default) will be hidden. Hence, if the user requests that the
old name is not to be kept for the clone, we recreate this file on the tmpfs.
This should be all that is required to restore the exact behaviour we would get
with a normal clone.
NOTE: If the container is rebooted all changes made to it are lost. This is not
easy to prevent since each reboot remounts the rootfs again.
Signed-off-by: Christian Brauner <cbrauner@suse.de>
This is a minimal commit which makes the function 'do_restore()' static
as it is not used anywhere else in the code. This also removes a
trailing space my editor complained about.
Signed-off-by: Adrian Reber <areber@redhat.com>
Shortly after CRIU 2.3 has been released a patch has been added to skip
in-flight TCP connections. In-flight connections are not completely
established connections (SYN, SYN-ACK). Skipping in-flight TCP
connections means that the client has to re-initiate the connection
establishment.
This patch stores the CRIU version detected during version check, so
that during dump/checkpoint options can be dynamically enabled depending
on the available CRIU version.
v2:
* use the newly introduced criu version interface
* add an option to disable skipping in-flight connections
Signed-off-by: Adrian Reber <areber@redhat.com>
The commit "c/r: add support for CRIU's --action-script" breaks
lxc-checkpoint on the command-line. It produces errors like:
sh: $'\260\366\b\001': command not found
and then it fails. src/lxc/criu.c expects migrate_opts->action_script to
be either NULL, then it is ignored, or to actually contain the name of
an action scripts.
As the struct migrate_opts has not static storage is has to be explicitly
initialized or the value of the structure's members is indeterminate.
Signed-off-by: Adrian Reber <areber@redhat.com>
- If version != NULL criu_version_ok() stores the detected criu version in
version. Allocates memory for version which must be freed by caller.
- If version == NULL criu_version_ok() will return true when the version
matches, false in all other cases.
Signed-off-by: Christian Brauner <cbrauner@suse.de>
First, we're doing this so long a there is any cgroup config item -
even if no devices ones. Then if devices is not available we fail.
This was leading to Rob E's mysterious startup failures.
Secondly, we're not even using this info. The user was removed
awhile back.
Signed-off-by: Serge Hallyn <serge@hallyn.com>
The regression was introduced by commit
3c39b0b7a2 which makes it possible to
create working stretch containers by forcinig `init` to be in the
included package list.
However, `init` didn't exit before jessie, so now for wheezy we
explicitly include `sysvinit`; sysvinit on wheezy is essential,
so it would already be included anyway.
Signed-off-by: Antonio Terceiro <terceiro@debian.org>
Newer versions of Android (5.0+, aka API Level 21+) include mntent.h,
which declares setmntent and endmntent. This hits an edge
case with the preprocessor checks in lxcmntent.h because HAVE_SETMNTENT
and HAVE_ENDMNTENT are both defined (in Bionic's mntent.h), but conf.c
always includes lxcmntent.h on Bionic! As a result, we get compiler
warnings of implicit function declarations for setmntent endmntent.
This patch always includes setmntent/endmntent/hasmntopt function
declarations on Bionic, which gets rid of these warnings.
Signed-off-by: Preetam D'Souza <preetamjdsouza@gmail.com>
A while ago cgroup modes were introduced to CRIU, which slightly changed
the behavior w.r.t. cgroups under the hood. What we're really after is
criu's --full mode, i.e. even if a particular cgroup directory exists
(in particular /lxc/$container[-$number] will, since we create it), we
should restore perms on that cgroup.
Things worked just fine for actual properties (except "special" properties
as criu refers to them, which I've just sent a patch for) because liblxc
creates no subdirectories, just the TLD.
Signed-off-by: Tycho Andersen <tycho.andersen@canonical.com>
The profile already contains
mount options=(rw, make-slave) -> **,
Which allows going through all mountpoints with make-slave,
so it seems to make sense to also allow the directly
recursive variant with "make-rslave".
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
Acked-by: Serge E. Hallyn <serge.hallyn@ubuntu.com>
Properly list all of the states and the right apparmor stanza for them,
then comment them all as actually enabling this would currently let the
user bypass apparmor entirely.
Signed-off-by: Stéphane Graber <stgraber@ubuntu.com>
Remove unnecessary shell wrap around job start.
Force foreground execution to allow job monitoring and control.
Signed-off-by Andrey Repin <anrdaemon@yandex.ru>
Bind-mounts aren't harmful in containers, so long as they're not used to
bypass MAC policies.
This change allows bind-mounting of any path which isn't a dangerous
filesystem that's otherwise blocked by apparmor.
This also allows switching paths {r}shared or {r}private.
Signed-off-by: Stéphane Graber <stgraber@ubuntu.com>
init 1.34 is not "Essential" anymore, in order to make it not required
on minimal chroots, docker containers, etc. Because of that we now need
to manually include it on systems that are expected to boot.
Signed-off-by: Antonio Terceiro <terceiro@debian.org>
If you specify an interpreter path with "-I" or "--interpreter-path",
the architecture of the debian container can differ from the one of
the host.
Before creating the container, binfmt must be configured on the host:
the script checks the name of the interpreter in /proc/sys/fs/binfmt_misc/
to know where to install it in the container.
To create a MIPS container on an x86_64 host:
$ cat /proc/sys/fs/binfmt_misc/qemu-mips
enabled
interpreter //qemu-mips
flags: OC
offset 0
magic 7f454c4601020100000000000000000000020008
mask ffffffffffffff00fffffffffffffffffffeffff
$ sudo lxc-create -n virtmips-stretch -t debian -- \
--arch=mips \
--interpreter-path=./mips-linux-user/qemu-mips \
--mirror=http://ftp.debian.org/debian \
--release=stretch
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
There is container-getty.service with OL7.2 systemd, it
is also used for managing the getty service, use that
instead and not manually create it.
Signed-off-by: Thomas Tanaka <thomas.tanaka@oracle.com>
When LXC is configured with --enable-rpath, I expect Python bindings
to be able to find the library in a non-standard location, just like
LXC command-line tools.
Signed-off-by: Aleksandr Mezin <mezin.alexander@gmail.com>
