IBM S/390 Reference Manual page 23

PR/SM Dynamic Storage Reconfiguration (DSR1).
The ability for an active logical partition to acquire pro-
cessor storage from another, adjacent logical partition.
This facility can be invoked via an operator command
entered at the system console, or automatically through
the PR/SM Automatic Reconfiguration Facility available
on selected processor models.
PR/SM Dynamic Storage Reconfiguration (DSR2).
Central storage (CS) can be reconfigured from one logi-
cal partition to another, even if the partitions are not
contiguous. Granularity is in multiples of 4 MB if CS < 2
GB, or multiples of 16 MB if CS = 2 GB.
PR/SM LPAR definitions retained.
Retains logical partition definitions (for example, parti-
tion weights, dedicated/shared status, capping) across
power-on resets even when a new I/O configuration data
set (IOCDS) is used. This reduces the manual time and
effort otherwise required to redefine these parameters.
PR/SM LPAR time management reporting.
This processor facility, along with RMF™ support, pro-
vides improved LPAR processor utilization reporting.
The RMF Partition Data Report provides total and effec-
tive partition dispatch times. This data enables custom-
ers to quantify the low utilization effect and do more
accurate capacity planning.
PR/SM LPAR preferred path.
Supports the availability to designate (in the IOCP) a
preferred channel path for I/O operations, when in
PR/SM LPAR mode (this function existed previously only
in basic mode). Always attempting new I/O requests to
the same path first can improve throughput for IBM
3480/3490 tape subsystems. Most installed tape sub-
systems have two control unit functions, each with its
own buffer storage and channel attachments, and each
42
with access to all tape drives. The subsystem employs
an algorithm that attempts to keep buffers for a given
drive in the control unit to which most I/O requests for
that drive are directed, avoiding internal overhead that
otherwise occurs.
PR/SM VARY CP ON/OFF.
Physical processors can be varied on/off line to a logical parti-
tion by operator command. This facility complements the con-
current central processor/TCM maintenance capability.
PR/SM Resource Capping.
The ability to limit the maximum amount of central pro-
cessor resource available to a given logical partition.
PR/SM up to 15 LPARs.
Single image processors can be configured to support
up to fifteen PR/SM logical partitions. This can be useful
in consolidating workloads onto fewer machines, and in
providing additional hot standby or test partitions. The
Multiple Image Facility (MIF) allows sharing of ESCON
channels and Internal Disk across multiple partitions,
thus controlling that aspect of cost.
RAID-5 Disk Array.
All Internal Disks are protected by RAID-5 disk array
technology to provide the capability to concurrently main-
tain HDDs (hot swap). By configuring a RAID-5 disk array,
one disk hardware failure per array does not affect pro-
cessing application. In the event of a disk failure, data is
still available using the parity recovery from another disks
in array. Maintenance can be performed concurrently.
System Assist Processor reassignment.
If the System Assist Processor (SAP) fails, a functioning
CP is quiesced and dynamically reassigned as an SAP
to maintain continuity of I/O processing. This capability
can eliminate an unplanned outage and permits the
repair to be deferred to a more convenient time.
47
43