Support checkpoint/restore with pods

This adds support to checkpoint containers out of pods and restore
container into pods.

It is only possible to restore a container into a pod if it has been
checkpointed out of pod. It is also not possible to restore a non pod
container into a pod.

The main reason this does not work is the PID namespace. If a non pod
container is being restored in a pod with a shared PID namespace, at
least one process in the restored container uses PID 1 which is already
in use by the infrastructure container. If someone tries to restore
container from a pod with a shared PID namespace without a shared PID
namespace it will also fail because the resulting PID namespace will not
have a PID 1.

Signed-off-by: Adrian Reber <areber@redhat.com>
This commit is contained in:
Adrian Reber
2021-07-12 11:43:45 +00:00
committed by Adrian Reber
parent 3375cbb198
commit eb94467780
12 changed files with 294 additions and 5 deletions

View File

@ -780,6 +780,16 @@ type ContainerCheckpointOptions struct {
// Compression tells the API which compression to use for
// the exported checkpoint archive.
Compression archive.Compression
// If Pod is set the container should be restored into the
// given Pod. If Pod is empty it is a restore without a Pod.
// Restoring a non Pod container into a Pod or a Pod container
// without a Pod is theoretically possible, but will
// probably not work if a PID namespace is shared.
// A shared PID namespace means that a Pod container has PID 1
// in the infrastructure container, but without the infrastructure
// container no PID 1 will be in the namespace and that is not
// possible.
Pod string
}
// Checkpoint checkpoints a container
@ -811,7 +821,11 @@ func (c *Container) Checkpoint(ctx context.Context, options ContainerCheckpointO
// Restore restores a container
func (c *Container) Restore(ctx context.Context, options ContainerCheckpointOptions) error {
logrus.Debugf("Trying to restore container %s", c.ID())
if options.Pod == "" {
logrus.Debugf("Trying to restore container %s", c.ID())
} else {
logrus.Debugf("Trying to restore container %s into pod %s", c.ID(), options.Pod)
}
if !c.batched {
c.lock.Lock()
defer c.lock.Unlock()