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

Planning Database Security and Recovery
HP NonStop SQL/MX Installation and Management Guide—523723-004
5-14
Guidelines for Creating and Managing Online
Dumps
•
The volume or volumes containing the audit trails must have sufficient free space
to accommodate the extents required for the number of audit-trail files. If there is
insufficient space to create a new audit-trail file, transactions can be suspended
while the TMF operator dumps older audit-trail files to tape and frees enough
space to continue.
•
The amount of audit-trail data generated can vary, depending on the setting of the
AUDITCOMPRESS attribute for the audited database files. Using
AUDITCOMPRESS saves system resources for update operations. Using NO
AUDITCOMPRESS enables you to read TMF audit-trail files with complete before
and after images. AUDITCOMPRESS is the default for audited tables.
For additional information on TMF configuration parameters and protection methods,
see the TMF Planning and Configuration Guide.
Guidelines for Creating and Managing Online Dumps
Correct handling of online dumps is essential for effective functioning of file recovery
protection. TMF does not determine a schedule for updating online dumps. You must
decide on an online dump schedule that satisfies the needs of your business
operations. You can make online dumps without stopping your applications.
When working with online dumps, consider these guidelines:
•
When you create an online dump, to prevent it from being purged when a database
is dropped, ensure that the online dump contains partitions from multiple database
tables. You should never create an online dump that contains all the partitions of a
single table and no partitions from any other table. If you do not follow this practice,
you will be unable to use TMF to recover accidentally dropped database tables.
•
You can send online dumps to disk or tape. Dumping to tape uses one tape drive
completely and some system resources. You might not want to schedule online
dumps and backups (see Using Backup and Restore 2 for Database Recovery on
page 5-27) at the same time or during peak hours of application processing.
•
You should coordinate online dumps with application activity. For example, if your
site performs a series of batch processing or weekly updates at a particular time
each week, you should follow those operations with online dumps of database
objects. Therefore, if a file recovery is necessary, your online dumps already reflect
the batch updates. TMF would need to apply only those database changes that
occurred after online dumps were taken.
•
When you create a new table and want to provide file recovery protection for it,
make an initial online dump of the table after creation.
•
Certain DDL and statements and utility commands invalidate previous online
dumps. For instance, whenever you load tables, update metadata, create new
indexes, partition tables or indexes, or restructure or move the database, you
should always make new online dumps to ensure the new status of the database is
recorded correctly. See Operations That Impact TMF Online Dumps on page 5-15.