HP-UX Workload Manager User's Guide

Configuring WLM
Tuning the metrics and the SLOs
Chapter 5 219
tune {
distribute_excess = 1;
}
Only workloads with active SLOs receive the excess CPU resources.
This distribution is based on balancing the weight-to-allocation ratios for
the workloads. These ratios are discussed in “Weighting a group so it
gets more CPU resources (optional)” on page 178.
The distribution is subject to the group CPU maximum values specified
by the gmaxcpu keyword. If the excess CPU resources are not fully
distributed because of group CPU maximum values, the excess is given
to the OTHERS group as usual—even if giving OTHERS this excess places it
over its gmaxcpu value.
The default value for distribute_excess is 0, in which case the OTHERS
group is given any excess CPU resources.
Refining granularity of CPU (and memory) allocation
by increasing shares per core (optional)
By default, allocations to FSS groups are based on 100 shares per core.
The minimum allocation to groups with inactive SLOs is one share or 1%
of CPU resources. If you are using WLM memory management, the
group receives 1% of memory. You can refine the granularity for
minimum allocations by setting the extended_shares tunable to 1 (true)
in a global tune structure:
tune {
extended_shares = 1;
}
In this way, CPU allocations are based on 1000 shares per core
(granularity of 0.001). This requires that you use absolute CPU units,
which is also recommended (set absolute_cpu_units to 1, as explained
in “Using absolute CPU units” on page 217). Minimum allocations of
memory are also based on 1000 shares instead of 100.
With extended_shares set to 1, the minimum allocation of CPU or
memory resources for a group with no active SLOs (this assumes
transient_groups is disabled) decreases from 1/100 core or 1% to
2/1000 core or 0.2%. Granularity for minimum allocations can be based
on 0.1% increases (for example, 0.3%, 0.4%, 0.5%). With