Chapter 1. Understanding OpenShift sandboxed containers

OpenShift sandboxed containers support for Red Hat OpenShift Container Platform provides you with built-in support for running Kata Containers as an additional optional runtime. The new runtime supports containers in dedicated virtual machines (VMs), providing improved workload isolation. This is particularly useful for performing the following tasks:

Run privileged or untrusted workloads

OpenShift sandboxed containers (OSC) makes it possible to safely run workloads that require specific privileges, without having to risk compromising cluster nodes by running privileged containers. Workloads that require special privileges include the following:

  • Workloads that require special capabilities from the kernel, beyond the default ones granted by standard container runtimes such as CRI-O, for example to access low-level networking features.
  • Workloads that need elevated root privileges, for example to access a specific physical device. With OpenShift sandboxed containers, it is possible to pass only a specific device through to the VM, ensuring that the workload cannot access or misconfigure the rest of the system.
  • Workloads for installing or using set-uid root binaries. These binaries grant special privileges and, as such, can present a security risk. With OpenShift sandboxed containers, additional privileges are restricted to the virtual machines, and grant no special access to the cluster nodes.

Some workloads may require privileges specifically for configuring the cluster nodes. Such workloads should still use privileged containers, because running on a virtual machine would prevent them from functioning.

Ensure kernel isolation for each workload
OpenShift sandboxed containers supports workloads that require custom kernel tuning (such as sysctl, scheduler changes, or cache tuning) and the creation of custom kernel modules (such as out of tree or special arguments).
Share the same workload across tenants
OpenShift sandboxed containers enables you to support multiple users (tenants) from different organizations sharing the same OpenShift cluster. The system also lets you run third-party workloads from multiple vendors, such as container network functions (CNFs) and enterprise applications. Third-party CNFs, for example, may not want their custom settings interfering with packet tuning or with sysctl variables set by other applications. Running inside a completely isolated kernel is helpful in preventing "noisy neighbor" configuration problems.
Ensure proper isolation and sandboxing for testing software
You can use OpenShift sandboxed containers to run a containerized workload with known vulnerabilities or to handle an issue in a legacy application. This isolation also enables administrators to give developers administrative control over pods, which is useful when the developer wants to test or validate configurations beyond those an administrator would typically grant. Administrators can, for example, safely and securely delegate kernel packet filtering (eBPF) to developers. Kernel packet filtering requires CAP_ADMIN or CAP_BPF privileges, and is therefore not allowed under a standard CRI-O configuration, as this would grant access to every process on the Container Host worker node. Similarly, administrators can grant access to intrusive tools such as SystemTap, or support the loading of custom kernel modules during their development.
Ensure default resource containment through VM boundaries
By default, resources such as CPU, memory, storage, or networking are managed in a more robust and secure way in OpenShift sandboxed containers. Since OpenShift sandboxed containers are deployed on VMs, additional layers of isolation and security give a finer-grained access control to the resource. For example, an errant container will not be able to allocate more memory than is available to the VM. Conversely, a container that needs dedicated access to a network card or to a disk can take complete control over that device without getting any access to other devices.

1.1. OpenShift sandboxed containers supported platforms

You can install OpenShift sandboxed containers on a bare-metal server or on an Amazon Web Services (AWS) bare-metal instance. Bare-metal instances offered by other cloud providers are not supported.

Red Hat Enterprise Linux CoreOS (RHCOS) is the only supported operating system for OpenShift sandboxed containers.

OpenShift sandboxed containers 1.5 is compatible with all supported versions of Red Hat OpenShift Container Platform. See the Red Hat OpenShift Container Platform Platform Life Cycle Policy for details.

1.2. OpenShift sandboxed containers common terms

The following terms are used throughout the documentation.

Sandbox

A sandbox is an isolated environment where programs can run. In a sandbox, you can run untested or untrusted programs without risking harm to the host machine or the operating system.

In the context of OpenShift sandboxed containers, sandboxing is achieved by running workloads in a different kernel using virtualization, providing enhanced control over the interactions between multiple workloads that run on the same host.

Pod

A pod is a construct that is inherited from Kubernetes and OpenShift Container Platform. It represents resources where containers can be deployed. Containers run inside of pods, and pods are used to specify resources that can be shared between multiple containers.

In the context of OpenShift sandboxed containers, a pod is implemented as a virtual machine. Several containers can run in the same pod on the same virtual machine.

OpenShift sandboxed containers Operator

An Operator is a software component that automates operations, which are actions that a human operator could do on the system.

The OpenShift sandboxed containers Operator is tasked with managing the lifecycle of sandboxed containers on a cluster. You can use the OpenShift sandboxed containers Operator to perform tasks such as the installation and removal of sandboxed containers, software updates, and status monitoring.

Kata Containers
Kata Containers is a core upstream project that is used to build OpenShift sandboxed containers. OpenShift sandboxed containers integrate Kata Containers with OpenShift Container Platform.
KataConfig
KataConfig objects represent configurations of sandboxed containers. They store information about the state of the cluster, such as the nodes on which the software is deployed.
Runtime class
A RuntimeClass object describes which runtime can be used to run a given workload. A runtime class that is named kata is installed and deployed by the OpenShift sandboxed containers Operator. The runtime class contains information about the runtime that describes resources that the runtime needs to operate, such as the pod overhead.
Peer pod
A peer pod in OpenShift sandboxed containers extends the concept of a standard pod. Unlike a standard sandboxed container, where the virtual machine is created on the worker node itself, in a peer pod, the virtual machine is created through a remote hypervisor using any supported hypervisor or cloud provider API. The peer pod acts as a regular pod on the worker node, with its corresponding VM running elsewhere. The remote location of the VM is transparent to the user and is specified by the runtime class in the pod specification. The peer pod design circumvents the need for nested virtualization.

1.3. OpenShift sandboxed containers workload management

OpenShift sandboxed containers provides the following features for enhancing workload management and allocation:

1.3.1. OpenShift sandboxed containers building blocks

The OpenShift sandboxed containers Operator encapsulates all of the components from Kata containers. It manages installation, lifecycle, and configuration tasks.

The OpenShift sandboxed containers Operator is packaged in the Operator bundle format as two container images. The bundle image contains metadata and is required to make the operator OLM-ready. The second container image contains the actual controller that monitors and manages the KataConfig resource.

1.3.2. RHCOS extensions

The OpenShift sandboxed containers Operator is based on the Red Hat Enterprise Linux CoreOS (RHCOS) extensions concept. RHCOS extensions are a mechanism to install optional OpenShift Container Platform software. The OpenShift sandboxed containers Operator uses this mechanism to deploy sandboxed containers on a cluster.

The sandboxed containers RHCOS extension contains RPMs for Kata, QEMU, and its dependencies. You can enable them by using the MachineConfig resources that the Machine Config Operator provides.

Additional resources

1.3.3. OpenShift Virtualization and OpenShift sandboxed containers

You can use OpenShift sandboxed containers on clusters with OpenShift Virtualization.

To run OpenShift Virtualization and OpenShift sandboxed containers at the same time, you must enable VMs to migrate, so that they do not block node reboots. Configure the following parameters on your VM:

  • Use ocs-storagecluster-ceph-rbd as the storage class.
  • Set the evictionStrategy parameter to LiveMigrate in the VM.

1.4. Understanding compliance and risk management

OpenShift Container Platform is designed for FIPS. When running Red Hat Enterprise Linux (RHEL) or Red Hat Enterprise Linux CoreOS (RHCOS) booted in FIPS mode, OpenShift Container Platform core components use the RHEL cryptographic libraries that have been submitted to NIST for FIPS 140-2/140-3 Validation on only the x86_64, ppc64le, and s390x architectures.

For more information about the NIST validation program, see Cryptographic Module Validation Program. For the latest NIST status for the individual versions of RHEL cryptographic libraries that have been submitted for validation, see Compliance Activities and Government Standards.

OpenShift sandboxed containers can be used on FIPS enabled clusters.

When running in FIPS mode, OpenShift sandboxed containers components, VMs, and VM images are adapted to comply with FIPS.

Note

FIPS compliance for OpenShift sandboxed containers only applies to the kata runtime class. The new peer pods runtime class kata-remote is not yet fully supported, and has not been tested for FIPS compliance.

FIPS compliance is one of the most critical components required in highly secure environments, to ensure that only supported cryptographic technologies are allowed on nodes.

Important

The use of FIPS Validated / Modules in Process cryptographic libraries is only supported on OpenShift Container Platform deployments on the x86_64 architecture.

To understand Red Hat’s view of OpenShift Container Platform compliance frameworks, refer to the Risk Management and Regulatory Readiness chapter of the OpenShift Security Guide Book.