User Namespaces

Rootless containers uses user_namespaces(7)(UserNS) for emulating fake privileges that are enough to create containers.

e.g. map UID 1000 to pseudo-root UID 0 in the UserNS:

$ whoami
$ id -u
$ unshare --user --map-root-user
# cat /proc/self/uid_map 
         0       1000          1
# cat /proc/self/gid_map 
         0       1000          1
# id -u

The pseudo-root user gains capabilities such as CAP_SYS_ADMIN and CAP_NET_ADMIN inside UserNS to perform fake-privileged operations such as creating mount namespaces, network namespaces, and creating TAP devices.

However, this user does not gain any actual privilege that can interfere with other users' files and processes.

subuids & subgids, newuidmap & newgidmap

Just mapping the single pseudo-root UID/GID is not enough to run containers that require multiple UIDs and GIDs. To map multiple UIDs and GIDs, Rootless Containers uses SETUID binaries called newuidmap and newgidmap. These binaries read configuration from files named /etc/subuid and /etc/subgid.

In the following example, 65,536 subuids (100000-165535) are allocated for a user named “user1”.

$ cat /etc/subuid

These subuids are mapped in the UserNS as follows:

1000 0
100000 1
100001 2
165535 65536

The same applies to subgids defined in /etc/subgid.

Creating a user namespace with newuidmap and newgidmap binaries are slightly complicated.

RootlessKit provides a simple CLI that wraps these binaries:

$ rootlesskit bash
# cat /proc/self/uid_map 
         0       1000          1
         1     100000      65536
# cat /proc/self/gid_map 
         0       1000          1
         1     100000      65536