1 Introduction #

In the Linux kernel, the cpuset facility provides a mechanism for creating logical entities called “cpusets” that encompass definitions of CPUs and NUMA Memory Nodes (if NUMA is available). Cpusets constrain the CPU and Memory placement of a task to only the resources defined within that cpuset. These cpusets can then be arranged into a nested hierarchy visible in the “cpuset” virtual file system. Sets of tasks can be assigned to these cpusets to constrain the resources that they use. The tasks can be moved from one cpuset to another to use other resources defined in those other cpusets.

The cset command is a Python application that provides a command line front-end for the Linux cpusets functionality. Working with cpusets directly can be confusing and slightly complex. The cset tool hides that complexity behind an easy-to-use command line interface.

There are two distinct use cases for cset: the basic shielding use case and the “advanced” case of using raw set and proc subcommands. The basic shielding function is accessed with the shield subcommand and described in the next section. Using the raw set and proc subcommands allows one to set up arbitrarily complex cpusets and is described in Chapter 3, Full Featured Cpuset Manipulation Commands.

Note that in general, one either uses the shield subcommand or a combination of the set and proc subcommands. One rarely, if ever, uses all of these subcommands together. Doing so will likely become too confusing. Additionally, the shield subcommand sets up its required cpusets with exclusively marked CPUs. This can interfere with your cpuset strategy. If you find that you need more functionality for your strategy than shield provides, go ahead and transition to using set and proc exclusively. It is straightforward to implement what shield does with a few extra set and proc subcommands.