Skip to content

Commit 683a43f

Browse files
committed
Minor wording update
1 parent 58a4db7 commit 683a43f

1 file changed

Lines changed: 1 addition & 1 deletion

File tree

pages/docs/admin-docs/docs-security.md

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -20,7 +20,7 @@ As mentioned, there are several containerization system calls and functions whic
2020

2121
1. **Limit usage to root:** Only allow the root user (or users granted `sudo`) to run containers. This has the obvious limitation of not allowing arbitrary users the ability to run containers, nor does it allow users to run containers as themselves. Access to data, security data, and securing systems becomes difficult and perhaps impossible.
2222
* **Root owned daemon process:** Some container systems use a root owned daemon background process which manages the containers and spawns the jobs within the container. Implementations of this typically have an IPC control socket for communicating with this root owned daemon process and if you wish to allow trusted users to control the daemon, you must give them access to the control socket. This is the Docker model.
23-
* **SetUID:** Set UID is an "old school" UNIX method for running a particular program with escalated permission. While it is widely used due to it's legacy and POSIX requirement, it lacks the ability to manage fine grained control of what a process can and can not do; a SetUID root program runs as root with all capabilities that comes with root. For this reason, SetUID programs are traditional targets for hackers.
23+
* **SetUID:** Set UID is the "old school" UNIX method for running a particular program with escalated permission. While it is widely used due to it's legacy and POSIX requirement, it lacks the ability to manage fine grained control of what a process can and can not do; a SetUID root program runs as root with all capabilities that comes with root. For this reason, SetUID programs are traditional targets for hackers.
2424
* **User Namespace:** The Linux kernel's user namespace may allow a user to virtually become another user and run a limited set privileged system functions. Here the privilege escalation is managed via the Linux kernel which takes the onus off of the program. This is a new kernel feature and thus requires new kernels and not all distributions have equally adopted this technology.
2525
* **Capability Sets:** Linux handles permissions, access, and roles via capability sets. The root user has these capabilities automatically activated while non-privileged users typically do not have these capabilities enabled. You can enable and disable capabilities on a per process and per file basis (if allowed to do so).
2626

0 commit comments

Comments
 (0)