SCF Reference Manual for the Storage Subsystem (G06.28+, H06.05+, J06.03+)

FASTBULKWRITE { ON | OFF },
(effective for all 512-byte-sector disks) specifies whether the system is using Fast Cache Bulk
Writes.
Cache Bulk Write is done in Fast mode.ON
Cache Bulk Write is done in Traditional I/O mode.OFF
The default is OFF.
When ON, applications and utilities using bulk writes to unstructured files may have higher
throughput. Using FASTBULKWRITE ON can result in lost data in unstructured files if the CPU
running the primary disk process fails. After a CPU failure, EMS event 5052 will report
unstructured files that have become broken as a result of the failure, and applications will
receive error 59 attempting to write to such files.
FASTBULKWRITE can be changed online or offline.
NOTE: These restrictions apply to partitioned unstructured files on an XP storage array and
on H-series, J-series, and S-series internal disks:
• All extents must be a multiple of 14 pages.
• All partitions must have identical extent sizes and maximum extents, because partitioning
depends on each partition having the same size.
FORCED
specifies that the disk attribute values be changed without any interaction with the user. SCF
does not prompt for confirmation.
FSTCACHING { ON | OFF }
specifies that the free-space table (FST) for a disk is updated in memory (ON), that the FST is
updated on disk (OFF) or that DP2 decides whether to keep the FST in memory or on disk
(ENABLED). The default is OFF. FSTCACHING ON can increase performance. The FST is
always rebuilt from disk when the disk is started.
The system disk always runs with FSTCACHING ON regardless of the configured value.
FULLCHECKPOINTS { ENABLED | DISABLED | FORCED }
(nondirectory structured files only) specifies when to perform full-block checkpoints
Never use a full-block checkpoint to protect the data on the disk. Using DISABLED maximizes disk
performance, but risks corrupting data blocks on the disk.
DISABLED
Use a full-block checkpoint only when serial writes are not being performed. Use ENABLED on
disks that perform parallel writes and on nonmirrored disks. For more details, see the SERIALWRITES
attribute (Default.)
ENABLED
Always use full-block checkpoint for disk-write operations.FORCED
For mirrored volumes, FULLCHECKPOINTS is important only if the primary or mirror disk is not
up. In that situation, FORCED or ENABLED can protect the validity of data written to the disk.
When performing full-block checkpoints, the primary process checkpoints the data to the backup
process before writing to the disk. This action ensures that the data on the disk is not corrupted
even if a hardware freeze or processor halt occurs during a disk write.
Use FORCED or ENABLED unless data loss is not critical.
ADD Command 203