CGroup Namespace Operations

This skill helps you work with Linux CGroup namespaces for process isolation, security analysis, and understanding container boundaries.

What are CGroup Namespaces?

CGroup namespaces provide isolation of cgroup hierarchies for processes. They virtualize the view of the cgroup hierarchy so that processes within a namespace see only their own cgroup subtree, with their own cgroup appearing as the root.

Key points:

• Cgroup namespaces isolate the view of the hierarchy, not the resources themselves
• Resource control is still enforced by cgroup subsystems (CPU, memory, I/O)
• Processes see their own cgroup as the root of the hierarchy
• They cannot see or access cgroups outside their subtree

When to Use This Skill

Use this skill when you need to:

• Create or enter cgroup namespaces
• Enumerate cgroup namespaces on a system
• Check which namespace a process belongs to
• Investigate container isolation boundaries
• Analyze privilege escalation paths involving namespaces
• Understand process isolation in containerized environments

Creating CGroup Namespaces

Using unshare (CLI)

# Create a new cgroup namespace with isolated /proc view
sudo unshare -C --mount-proc /bin/bash

# Create with fork (prevents PID allocation errors)
sudo unshare -fC --mount-proc /bin/bash

Important: Use the

-f

Using Docker

docker run -ti --name ubuntu1 -v /usr:/ubuntu1 ubuntu bash

Docker containers automatically create cgroup namespaces for isolation.

Inspecting CGroup Namespaces

Check your current namespace

ls -l /proc/self/ns/cgroup
# Output: cgroup:[4026531835]

The number in brackets is your cgroup namespace identifier.

Find all cgroup namespaces on the system

sudo find /proc -maxdepth 3 -type l -name cgroup -exec readlink {} \; 2>/dev/null | sort -u

Find processes in a specific namespace

# Replace <ns-number> with the namespace ID
sudo find /proc -maxdepth 3 -type l -name cgroup -exec ls -l {} \; 2>/dev/null | grep <ns-number>

Entering CGroup Namespaces

# Enter another process's cgroup namespace
nsenter -C TARGET_PID --pid /bin/bash

Requirements:

• You must be root to enter another process's namespace
• You need a descriptor pointing to the namespace (like

  /proc/self/ns/cgroup

  )

Security Considerations

Privilege Escalation Context

CGroup namespaces can be relevant for privilege escalation because:

1. Namespace enumeration can reveal container boundaries and isolation gaps
2. Namespace entry (if you have root) allows you to inspect processes in different isolation contexts
3. Understanding cgroup views helps identify what resources a process can see and potentially manipulate

Common Investigation Patterns

1. Map namespace boundaries: Find all cgroup namespaces and their associated processes
2. Check isolation: Verify that containers are properly isolated from the host
3. Identify escape vectors: Look for processes that can enter namespaces they shouldn't
4. Analyze resource visibility: Understand what cgroup hierarchies processes can see

Troubleshooting

"Cannot allocate memory" error with unshare

Cause: The unshare process doesn't enter the new namespace, and when PID 1 exits, the namespace is cleaned up.

Solution: Use the

-f

sudo unshare -fC --mount-proc /bin/bash

Cannot enter namespace

Cause: You're not root, or you don't have a namespace descriptor.

Solution:

• Ensure you have root privileges
• Use a valid namespace path like

  /proc/<pid>/ns/cgroup

Quick Reference

sudo unshare -fC --mount-proc /bin/bash

ls -l /proc/self/ns/cgroup

sudo find /proc -maxdepth 3 -type l -name cgroup -exec readlink {} \; 2>/dev/null | sort -u

sudo find /proc -maxdepth 3 -type l -name cgroup -exec ls -l {} \; 2>/dev/null | grep <ns-id>

nsenter -C TARGET_PID --pid /bin/bash

Next Steps

After working with cgroup namespaces, you may want to:

• Investigate other namespace types (PID, mount, network, user)
• Analyze cgroup resource limits and controls
• Examine container escape techniques
• Review process isolation boundaries

For more information on cgroups themselves, see the cgroups documentation.