HP Integrity NonStop H-Series Guía de usuario Pagina 238

  • Descarga
  • Añadir a mis manuales
  • Imprimir
  • Pagina
    / 346
  • Tabla de contenidos
  • MARCADORES
  • Valorado. / 5. Basado en revisión del cliente
Vista de pagina 237
HP NonStop Shadowbase Command Definitions
Database Specification Command Descriptions
HP NonStop Shadowbase Command Reference Manual—785421-002
229
DELETES { ON }
{ OFF }
Initial Changed
Version Version Values Default Minimum Maximum Alter
2.000 ON OFF ON No
This indicates whether the source database deletes are to be applied to
the target. The default is ON.
Note: When the DELETES parameter is set to OFF, deletes are not
applied to the target, but insert undo events are applied.
DELUIDXDUP { ON }
{ OFF }
Initial Changed
Version Version Values Default Minimum Maximum Alter
3.964B OFF OFF ON No
When the target Consumer is applying an I/O event against a target file or
table with a unique alternate key file or index, it is possible for the base file
I/O to be successful yet the I/O to the alternate key file or index to fail with
a duplicate key error. A simple example illustrates the issue – assume a
uni-directional replication scenario where the target database is not initially
empty and is not synchronized with the source database (this is often
referred to as a “fuzzy replication” scenario):
Assume the source database table is initially empty, yet the target
database table has a record with primary key value of 1 and a unique
index value of 1. In other words, the two are not initially synchronized
when replication is enabled/started.
Now assume that a source application inserts a row into the source
table with primary key value of 2 and a unique index value of 1. This
source insert is successful.
When the replication engine applies that change into the target
database, the base table insert will be successful, yet the unique index
insert will fail with a duplicate key error (which backs out, or undoes,
the entire insert on the target). In the previous versions of HP
Shadowbase, this duplicate key error would be mistakenly viewed as a
duplicate key error on the primary key (not the unique index) and
would cause the INSERT event to be mapped into an UPDATE (if the
fuzzy replication DBS parameter UPDATEDUPLICATE is ON), and re-
applied. Executing this UPDATE will cause a “not found” condition (as
no record with a primary key value of 2 exists in the target table), and
then the UPDATE would be logged and discarded. In other words, the
Vista de pagina 237
1 2 ... 233 234 235 236 237 238 239 240 241 242 243 ... 345 346

Comentarios a estos manuales

Sin comentarios