When C/R was implemented, it was enough to call manager.Set to apply
limits and to move a task. Now .Set() and .Apply() have to be called
separately.
Fixes: 8a740d5391 ("libcontainer: cgroups: don't Set in Apply")
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
Instead of relying on version numbers it is possible to check if CRIU
actually supports certain features. This introduces an initial
implementation to check if CRIU and the underlying kernel actually
support dirty memory tracking for memory pre-dumping.
Upstream CRIU also supports the lazy-page migration feature check and
additional feature checks can be included in CRIU to reduce the version
number parsing. There are also certain CRIU features which depend on one
side on the CRIU version but also require certain kernel versions to
actually work. CRIU knows if it can do certain things on the kernel it
is running on and using the feature check RPC interface makes it easier
for runc to decide if the criu+kernel combination will support that
feature.
Feature checking was introduced with CRIU 1.8. Running with older CRIU
versions will ignore the feature check functionality and behave just
like it used to.
v2:
- Do not use reflection to compare requested and responded
features. Checking which feature is available is now hardcoded
and needs to be adapted for every new feature check. The code
is now much more readable and simpler.
v3:
- Move the variable criuFeat out of the linuxContainer struct,
as it is not container specific. Now it is a global variable.
Signed-off-by: Adrian Reber <areber@redhat.com>
Previously Host{U,G}ID only gave you the root mapping, which isn't very
useful if you are trying to do other things with the IDMaps.
Signed-off-by: Aleksa Sarai <asarai@suse.de>
The rootless cgroup manager acts as a noop for all set and apply
operations. It is just used for rootless setups. Currently this is far
too simple (we need to add opportunistic cgroup management), but is good
enough as a first-pass at a noop cgroup manager.
Signed-off-by: Aleksa Sarai <asarai@suse.de>
This enables the support for the rootless container mode. There are many
restrictions on what rootless containers can do, so many different runC
commands have been disabled:
* runc checkpoint
* runc events
* runc pause
* runc ps
* runc restore
* runc resume
* runc update
The following commands work:
* runc create
* runc delete
* runc exec
* runc kill
* runc list
* runc run
* runc spec
* runc state
In addition, any specification options that imply joining cgroups have
also been disabled. This is due to support for unprivileged subtree
management not being available from Linux upstream.
Signed-off-by: Aleksa Sarai <asarai@suse.de>
Effectively, !dumpable makes implementing rootless containers quite
hard, due to a bunch of different operations on /proc/self no longer
being possible without reordering everything.
!dumpable only really makes sense when you are switching between
different security contexts, which is only the case when we are joining
namespaces. Unfortunately this means that !dumpable will still have
issues in this instance, and it should only be necessary to set
!dumpable if we are not joining USER namespaces (new kernels have
protections that make !dumpable no longer necessary). But that's a topic
for another time.
This also includes code to unset and then re-set dumpable when doing the
USER namespace mappings. This should also be safe because in principle
processes in a container can't see us until after we fork into the PID
namespace (which happens after the user mapping).
In rootless containers, it is not possible to set a non-dumpable
process's /proc/self/oom_score_adj (it's owned by root and thus not
writeable). Thus, it needs to be set inside nsexec before we set
ourselves as non-dumpable.
Signed-off-by: Aleksa Sarai <asarai@suse.de>
If we try to pause a container on the system without freezer cgroups,
we can found that runc tries to open ./freezer.state. It is obviously wrong.
$ ./runc pause test
no such directory for freezer.state
$ echo FROZEN > freezer.state
$ ./runc pause test
container not running or created: paused
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
In container process's Init function, we use
fd + execFifoFilename to open exec fifo, so this
field in init config is never used.
Signed-off-by: Qiang Huang <h.huangqiang@huawei.com>
It should not be binded to container creation, for
example, runc restore needs to create a
libcontainer.Container, but it won't need exec fifo.
So create exec fifo when container is started or run,
where we really need it.
Signed-off-by: Qiang Huang <h.huangqiang@huawei.com>
CRIU gets pre-dump to complete iterative migration.
pre-dump saves process memory info only. And it need parent-path
to specify the former memory files.
This patch add pre-dump and parent-path arguments to runc checkpoint
Signed-off-by: Deng Guangxing <dengguangxing@huawei.com>
Signed-off-by: Adrian Reber <areber@redhat.com>
If we pass a file descriptor to the host filesystem while joining a
container, there is a race condition where a process inside the
container can ptrace(2) the joining process and stop it from closing its
file descriptor to the stateDirFd. Then the process can access the
*host* filesystem from that file descriptor. This was fixed in part by
5d93fed3d2 ("Set init processes as non-dumpable"), but that fix is
more of a hail-mary than an actual fix for the underlying issue.
To fix this, don't open or pass the stateDirFd to the init process
unless we're creating a new container. A proper fix for this would be to
remove the need for even passing around directory file descriptors
(which are quite dangerous in the context of mount namespaces).
There is still an issue with containers that have CAP_SYS_PTRACE and are
using the setns(2)-style of joining a container namespace. Currently I'm
not really sure how to fix it without rampant layer violation.
Fixes: CVE-2016-9962
Fixes: 5d93fed3d2 ("Set init processes as non-dumpable")
Signed-off-by: Aleksa Sarai <asarai@suse.de>
`HookState` struct should follow definition of `State` in runtime-spec:
* modify json name of `version` to `ociVersion`.
* Remove redundant `Rootfs` field as rootfs can be retrived from
`bundlePath/config.json`.
Signed-off-by: Zhang Wei <zhangwei555@huawei.com>
This implements {createTTY, detach} and all of the combinations and
negations of the two that were previously implemented. There are some
valid questions about out-of-OCI-scope topics like !createTTY and how
things should be handled (why do we dup the current stdio to the
process, and how is that not a security issue). However, these will be
dealt with in a separate patchset.
In order to allow for late console setup, split setupRootfs into the
"preparation" section where all of the mounts are created and the
"finalize" section where we pivot_root and set things as ro. In between
the two we can set up all of the console mountpoints and symlinks we
need.
We use two-stage synchronisation to ensures that when the syscalls are
reordered in a suboptimal way, an out-of-place read() on the parentPipe
will not gobble the ancilliary information.
This patch is part of the console rewrite patchset.
Signed-off-by: Aleksa Sarai <asarai@suse.de>
This allows a user to send a signal to all the processes in the
container within a single atomic action to avoid new processes being
forked off before the signal can be sent.
This is basically taking functionality that we already use being
`delete` and exposing it ok the `kill` command by adding a flag.
Signed-off-by: Michael Crosby <crosbymichael@gmail.com>
In user namespaces devices are bind-mounted from the host, so
we need to add them as external mounts for CRIU.
Reported-by: Ross Boucher <boucher@gmail.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
Depending on your SELinux setup, the order in which you join namespaces
can be important. In general, user namespaces should *always* be joined
and unshared first because then the other namespaces are correctly
pinned and you have the right priviliges within them. This also is very
useful for rootless containers, as well as older kernels that had
essentially broken unshare(2) and clone(2) implementations.
This also includes huge refactorings in how we spawn processes for
complicated reasons that I don't want to get into because it will make
me spiral into a cloud of rage. The reasoning is in the giant comment in
clone_parent. Have fun.
In addition, because we now create multiple children with CLONE_PARENT,
we cannot wait for them to SIGCHLD us in the case of a death. Thus, we
have to resort to having a child kindly send us their exit code before
they die. Hopefully this all works okay, but at this point there's not
much more than we can do.
Signed-off-by: Aleksa Sarai <asarai@suse.de>
This avoids us from running into cases where libcontainer thinks that a
particular namespace file is a different type, and makes it a fatal
error rather than causing broken functionality.
Signed-off-by: Aleksa Sarai <asarai@suse.de>
if a container state is running or created, the container.Pause()
method can set the state to pausing, and then paused.
this patch update the comment, so it can be consistent with the code.
Signed-off-by: Wang Long <long.wanglong@huawei.com>
1. According to docs of Cmd.Path and Cmd.Args from package "os/exec":
Path is the path of the command to run. Args holds command line
arguments, including the command as Args[0]. We have mixed usage
of args. In InitPath(), InitArgs only take arguments, in InitArgs(),
InitArgs including the command as Args[0]. This is confusing.
2. InitArgs() already have the ability to configure a LinuxFactory
with the provided absolute path to the init binary and arguements as
InitPath() does.
3. exec.Command() will take care of serching executable path.
4. The default "/proc/self/exe" instead of os.Args[0] is passed to
InitArgs in order to allow relative path for the runC binary.
Signed-off-by: Yang Hongyang <imhy.yang@gmail.com>
This removes the use of a signal handler and SIGCONT to signal the init
process to exec the users process.
Signed-off-by: Michael Crosby <crosbymichael@gmail.com>
This is the inital port of the libcontainer.Error to added a cause to
all the existing error messages. Going forward, when an error can be
wrapped because it is not being checked at the higher levels for
something like `os.IsNotExist` we can add more information to the error
message like cause and stack file/line information. This will help
higher level tools to know what cause a container start or operation to
fail.
Signed-off-by: Michael Crosby <crosbymichael@gmail.com>
No substantial code change.
Note that some style errors reported by `golint` are not fixed due to possible compatibility issues.
Signed-off-by: Akihiro Suda <suda.kyoto@gmail.com>
This updates runc and libcontainer to handle rlimits per process and set
them correctly for the container.
Signed-off-by: Michael Crosby <crosbymichael@gmail.com>
This commit adds support to libcontainer to allow caps, no new privs,
apparmor, and selinux process label to the process struct so that it can
be used together of override the base settings on the container config
per individual process.
Signed-off-by: Michael Crosby <crosbymichael@gmail.com>
currentState() always adds all possible namespaces to the state,
regardless of whether they are supported.
If orderNamespacePaths detects an unsupported namespace, an error is
returned that results in initialization failure.
Fix this by only adding paths of supported namespaces to the state.
Signed-off-by: Ido Yariv <ido@wizery.com>
An init process can join other namespaces (pidns, ipc etc.). This leverages
C code defined in nsenter package to spawn a process with correct namespaces
and clone if necessary.
This moves all setns and cloneflags related code to nsenter layer, which mean
that we dont use Go os/exec to create process with cloneflags and set
uid/gid_map or setgroups anymore. The necessary data is passed from Go to C
using a netlink binary-encoding format.
With this change, setns and init processes are almost the same, which brings
some opportunity for refactoring.
Signed-off-by: Daniel, Dao Quang Minh <dqminh89@gmail.com>
[mickael.laventure@docker.com: adapted to apply on master @ d97d5e]
Signed-off-by: Kenfe-Mickael Laventure <mickael.laventure@docker.com>
This adds orderNamespacePaths to get correct order of namespaces for the
bootstrap program to join.
Signed-off-by: Daniel, Dao Quang Minh <dqminh89@gmail.com>
Create a unique session key name for every container. Use the pattern
_ses.<postfix> with postfix being the container's Id.
This patch does not prevent containers from joining each other's session
keyring.
Signed-off-by: Stefan Berger <stefanb@linux.vnet.ibm.com>
Docker uses Prestart hooks to call a libnetwork hook to create
network devices and set addesses and routes.
Signed-off-by: Andrew Vagin <avagin@virtuozzo.com>
This options is set a namespace mask which will not be dumped and restored.
For example, we are going to use this option to restore network
for docker containers. CRIU will create a network namespace and
call a libnetwork hook to restore network devices, addresses and routes.
Signed-off-by: Andrew Vagin <avagin@virtuozzo.com>
We don't need a CreatedTime method on the container because it's not
part of the interface and can be received via the state. We also do not
need to call it CreateTime because the type of this field is time.Time
so we know its time.
Signed-off-by: Michael Crosby <crosbymichael@gmail.com>
Marshall the raw objects for the sync pipes so that no new line chars
are left behind in the pipe causing errors.
Signed-off-by: Michael Crosby <crosbymichael@gmail.com>
There were issues where a process could die before pausing completed
leaving the container in an inconsistent state and unable to be
destoryed. This makes sure that if the container is paused and the
process is dead it will unfreeze the cgroup before removing them.
Signed-off-by: Michael Crosby <crosbymichael@gmail.com>
It may be desirable to receive memory pressure levels notifications
before the container depletes all memory. This may be useful for
handling cases where the system thrashes when reaching the container's
memory limits.
Signed-off-by: Ido Yariv <ido@wizery.com>
Signed-off-by: Michael Crosby <crosbymichael@gmail.com>
Add state status() method
Signed-off-by: Michael Crosby <crosbymichael@gmail.com>
Allow multiple checkpoint on restore
Signed-off-by: Michael Crosby <crosbymichael@gmail.com>
Handle leave-running state
Signed-off-by: Michael Crosby <crosbymichael@gmail.com>
Fix state transitions for inprocess
Because the tests use libcontainer in process between the various states
we need to ensure that that usecase works as well as the out of process
one.
Signed-off-by: Michael Crosby <crosbymichael@gmail.com>
Remove isDestroyed method
Signed-off-by: Michael Crosby <crosbymichael@gmail.com>
Handling Pausing from freezer state
Signed-off-by: Rajasekaran <rajasec79@gmail.com>
freezer status
Signed-off-by: Rajasekaran <rajasec79@gmail.com>
Fixing review comments
Signed-off-by: Rajasekaran <rajasec79@gmail.com>
Added comment when freezer not available
Signed-off-by: Rajasekaran <rajasec79@gmail.com>
Signed-off-by: Michael Crosby <crosbymichael@gmail.com>
Conflicts:
libcontainer/container_linux.go
Change checkFreezer logic to isPaused()
Signed-off-by: Michael Crosby <crosbymichael@gmail.com>
Remove state base and factor out destroy func
Signed-off-by: Michael Crosby <crosbymichael@gmail.com>
Add unit test for state transitions
Signed-off-by: Michael Crosby <crosbymichael@gmail.com>
This allows us to distinguish cases where a container
needs to just join the paths or also additionally
set cgroups settings. This will help in implementing
cgroupsPath support in the spec.
Signed-off-by: Mrunal Patel <mrunalp@gmail.com>
replace passing of pid and console path via environment variable with passing
them with netlink message via an established pipe.
this change requires us to set _LIBCONTAINER_INITTYPE and
_LIBCONTAINER_INITPIPE as the env environment of the bootstrap process as we
only send the bootstrap data for setns process right now. When init and setns
bootstrap process are unified (i.e., init use nsexec instead of Go to clone new
process), we can remove _LIBCONTAINER_INITTYPE.
Note:
- we read nlmsghdr first before reading the content so we can get the total
length of the payload and allocate buffer properly instead of allocating
one large buffer.
- check read bytes vs the wanted number. It's an error if we failed to read
the desired number of bytes from the pipe into the buffer.
Signed-off-by: Daniel, Dao Quang Minh <dqminh89@gmail.com>
add bootstrap data to setns process. If we have any bootstrap data then copy it
to the bootstrap process (i.e. nsexec) using the sync pipe. This will allow us
to eventually replace environment variable usage with more structured data
to setup namespaces, write pid/gid map, setgroup etc.
Signed-off-by: Daniel, Dao Quang Minh <dqminh89@gmail.com>
When starting and quering for pids a container can start and exit before
this is set. So set the opts after the process is started and while
libcontainer still has the container's process blocking on the pipe.
Signed-off-by: Michael Crosby <crosbymichael@gmail.com>
While testing different versions of criu it helps to know which
criu binary with which options is currently used. Therefore additional
debug output to display these information is added.
v2: increase readability of printed out criu options
Signed-off-by: Adrian Reber <adrian@lisas.de>
Here are two reasons:
* If we use systemd, we need to ask it to create cgroups
* If a container is restored with another ID, we need to
change paths to cgroups.
Signed-off-by: Andrey Vagin <avagin@openvz.org>
This adds a `Signal()` method to the container interface so that the
initial process can be signaled after a Load or operation. It also
implements signaling the init process from a nonChildProcess.
Signed-off-by: Michael Crosby <crosbymichael@gmail.com>
A boolean field named GidMappingsEnableSetgroups was added to
SysProcAttr in Go1.5. This field determines the value of the process's
setgroups proc entry.
Since the default is to set the entry to 'deny', calling setgroups will
fail on systems running kernels 3.19+.
Set GidMappingsEnableSetgroups to true so setgroups wont be set to
'deny'.
Signed-off-by: Ido Yariv <ido@wizery.com>
Actually cgroup mounts are bind-mounts, so they should be
handled by the same way.
Reported-by: Ross Boucher <rboucher@gmail.com>
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Qiang Huang
Mrunal Patel
Andrei Vagin
Adrian Reber
Harshal Patil
Aleksa Sarai
Michael Crosby
Mrunal Patel
Deng Guangxing
Zhang Wei
Xianlu Bird
Wang Long
rajasec
Phil Estes
Yang Hongyang
Petar Petrov
Akihiro Suda
George Lestaris
Peng Gao
Jessica Frazelle
Ido Yariv
Daniel, Dao Quang Minh
Stefan Berger
Mike Brown
Doug Davis
Jimmi Dyson
John Howard
Adrian Reber
Hui Kang
David Calavera
Andrey Vagin
boucher
mapk0y
unclejack