The system administrator can define up to 64 superclasses.
In addition, the following five superclasses are automatically created:
- Default superclass
- Is the default superclass and is always defined. All non-root processes that are not automatically assigned to a specific superclass are assigned to the Default superclass. Other processes can also be assigned to the Default superclass by providing specific assignment rules.
- System superclass
- Has all privileged (root) processes assigned to it if those processes are not assigned by rules to a specific class. This superclass also collects the memory pages belonging to kernel memory segments and kernel processes. Other processes can also be assigned to the System superclass by providing specific assignment rules for this superclass. This superclass has a memory minimum limit of 1% as the default.
- Shared superclass
- Receives the memory pages that are shared by processes in more than one superclass. This includes pages in shared memory regions and pages in files that are used by processes in more than one superclass (or in subclasses of different superclasses). Shared memory and files that are used by multiple processes that all belong to a single superclass (or subclasses of the same superclass) are associated with that superclass. Only when a process from a different superclass accesses the shared memory region or file are the pages placed in the Shared superclass. This superclass can have only physical memory shares and limits applied to it. It cannot have shares or limits for the other resource types, subclasses, or assignment rules specified. Whether a memory segment shared by processes in different subclasses of the same superclass is classified into the Shared subclass or remains in its original subclass depends on the value of the localshm attribute of the original subclass.
- Unclassified superclass
- Is a memory allocation for unclassified processes. The processes in existence
at the time that WLM is started are classified according to the assignment
rules of the WLM configuration being loaded. During this initial classification,
all the memory pages attached to each process are "charged" either to the
superclass that the process belongs to (when not shared, or shared by processes
in the same superclass), or to the Shared superclass when shared by
processes in different superclasses.
However, a few pages cannot be directly tied to any processes (and thus to any class) at the time of this classification, and this memory is charged to the Unclassified superclass. Most of this memory is correctly reclassified over time, when it is either accessed by a process, or released and reallocated to a process after WLM is started. There are no processes in the Unclassified superclass. This superclass can have physical memory shares and limits applied to it. It cannot have shares or limits for the other resource types, subclasses, or assignment rules specified.
- Unmanaged superclass
- A special superclass, named Unmanaged, is always defined. No processes are assigned to this class. This class accumulates the memory usage for all pinned pages in the system that are not managed by WLM. The CPU utilization for the waitproc processes is not accumulated in any class to prevent the system from appearing to be at 100% utilization. This superclass cannot have shares or limits for any resource types, subclasses, or specified assignment rules.