SQL/MX 2.x Installation and Management Guide (G06.24+, H06.03+)

Introduction to SQL/MX Database Management
HP NonStop SQL/MX Installation and Management Guide—523723-004
1-6
High Availability
When tables and indexes are partitioned across multiple disks, NonStop SQL/MX can
use a different process for each partition during query execution. This approach
reduces the time for scans and other set-oriented processing by a factor equivalent to
the number of partitions when there is no contention in the processor-controller-disk
path; that is, when every participating disk is primary to a different processor.
Parallel join operations are performed by the SQL/MX executor during query
processing.
Parallel index maintenance reduces the effect of multiple indexes on performance.
Each index on a table is automatically updated whenever a row is inserted into the
table and whenever a value is updated in any key column of the index. Multiple
indexes on a table can be updated in parallel by different disk processes or by the
same disk process servicing multiple requests.
Indexes are loaded in parallel. Parallel index loading speeds the loading of a
partitioned index by loading all partitions at the same time.
Parallel sorting is performed by the executor server processes (ESPs).
Parallel input-output operations are performed on multiple partitions by different disk
processes. A single disk process can also perform parallel I/O by buffering operations
in cache.
High Availability
These SQL/MX features help to ensure high availability for databases:
•
Online dumps using the TMFCOM DUMP FILES command, with complete support
for TMF file recovery to recover a database in case of a disaster.
•
Online database reorganization capabilities such as online partition moving,
partition splitting, and table and index row redistribution with concurrent read and
update capability. For more information, see Section 10, Reorganizing SQL/MX
Tables and Maintaining Data.
•
Parallel table reorganization (using FUP RELOAD) and index loads (using
POPULATE INDEX) to reduce the time required to load the object.
•
Automatic recompilation, which eliminates the need to terminate program
execution when changes in database structure or the environment make rebinding
necessary.
•
The ability to defer name resolution in SQL/MX statements until execution time