| administration
features |
| Datafile
Management: |
|
|
|
 |
New
feature allows for automatic management of datafiles when
you issue tablespace commands. By specifying a directory
location in the init.ora file (or by an 'alter system' command),
the system will automatically create and delete datafiles
in that location as necessary. A new addition to the drop
tablespace' clause ('drop tablespace…including contents
and datafiles') will drop all associated datafiles be they
automatically or manually managed. |
|
| Server
Parameter File (spfile): |
|
|
|
| Initially created
from a standard text parameter file, this new binary file is
created
and maintained on the machine housing the Oracle software.
On-line parameter changes made using 'alter system' commands
are stored here meaning that such updates can persist across
start-up and shutdown. Additionally you can start-up a database
from a client by referencing the spfile in a parameter file
on the client machine. |
|
| On-line Table Redefinition: |
|
|
|
| Using the new DBMS_REDEFINITION package you can completely redefine tables online. You can use the package to create a copy of the current table in its new incarnation and keep the copy table data updated in line with the current table if required. Then, when you are ready to complete the move, you use the package to transparently switch to the new table. |
|
| Automatic Undo Management |
|
|
|
| In an effort to reduce administration overheads, it is now possible to have the system manage it's own rollback segments (now known as "undo" segments) by specifying the undo tablespace and undo management type in the parameter file. The system-managed undo tablespace is essentially a locally managed tablespace with space managed by a bitmap in the file header. |
|
| Database Suspend/Resume |
|
|
|
| This feature allows you to suspend database activity by stopping I/O to all datafiles and queuing any further I/O requests. When ready, you can then resume database operations and the system will re-activate all pending I/O requests in the order they were queued. This is extremely useful for system backups that involve splitting mirrored disks. |
|
| Database Quiescence: |
|
|
|
| You place the database in a quiesced state in which only the SYS and SYSTEM users can perform transactions. This basically removes the need to stop the database and restart in restricted mode if you want to perform tasks that could interfere with normal usage. |
|
| Resumable
Space Allocation: |
|
|
|
|
This
excellent new feature allows you to resume large transactions
that have failed due to some space allocation issue such
as an object reaching its max extents or a user quota limit
exceeded. Suspended transactions are noted in the alert log
but a new trigger type (an 'after suspend' trigger) can be
used in conjunction with a new PL/SQL package to automatically
deal with correctable errors. |
|
|
|
 |
| Dynamic
SGA: |
|
|
|
| The SGA can now be resized without having to restart the database. By issuing 'alter system' commands you can resize the shared pool, large pool, buffer cache and log buffer. The buffer cache can only be dynamically resized if you are using the new DB_CACHE_SIZE parameter. You can also place a limit on the size of the SGA by using the SGA_MAX_SIZE initialisation parameter. |
|
| Multiple Block Size: |
|
|
|
|
|
| Tablespaces can be created with non-standard block sizes (i.e. block sizes other than that specified by the DB_BLOCK_SIZE initialisation parameter). To use this feature you must specify a sub-cache for that block size by using the appropriate DBA_(n)K_CACHE_SIZE initialisation parameter. |
|
| Automatic Segment Space Allocation: |
|
|
|
| The locally managed tablespace feature which was introduced in Oracle 8i has now been enhanced so that the system can automatically manage space within segments as well as space within the tablespace. This means that Oracle uses bitmaps to manage space within segments instead of using freelists, the traditional method. This means that you no longer need to tune the PCTUSED, FREELIST and FREELIST GROUPS segment attributes. |
|
| Default
Temporary Tablespace: |
|
|
|
| The system will automatically allocate new users an appropriate temporary tablespace without you needing to specify it explicitly in the 'create user' statement. This is achieved by specifying a pre-created temporary tablespace as default using a new extension of the 'create tablespace' and 'alter tablespace' commands. |
|
| Time-based
Logfile Switching: |
|
|
|
| By setting a new ARCHIVE_LAG_TARGET initialisation parameter you can cause logfile switches at timed intervals provided there is redo information within the current archive log. This feature is especially useful for standby databases as it limits the time between information being changed in the primary database and that information reaching the standby database through archive log shipping. |
|
| The Metadata API: |
|
|
|
|
A
new PL/SQL package, DBMS_METADATA, allows you to view metadata
about a schema object in the form of DDL statements. |
|
|
|
 |
| More
Archiving Destinations: |
|
|
|
| You
can now have 5-10 archive log destinations specified in the
parameter file. |
|
| Constraint
Enhancements: |
|
|
|
| Enhancements to the 'using index' clause of 'create table' and 'alter table' statements means you can specify the creation or use of a specific index when dealing with primary keys. You can also prevent indexes from being dropped when dropping constraints. |
|
| Database
Time Zone: |
|
|
|
| By specifying a time zone at database creation, the database can operate in a different time zone to the underlying operating system. Dates and times in transactions can also be resolved to the database time zone when applying data. |
|
| Monitoring
Index Usage: |
|
|
|
| By using the 'alter index … monitoring usage' statement you can see if a particular index is being used. Usage data is then viewed through the v$object_usage system view. |
|
| Dynamic Job Queue Processes: |
|
|
|
| These processes are now spawned dynamically through a single co-ordinating process. This means that only the necessary numbers of job queue processes exist at any one time allowing more efficient memory usage. |
|
| Fine
Grained Auditing: |
|
|
|
|
| You can now audit in much greater detail than before. For example, you can use auditing to determine the nature of the data that someone has selected from a table as well as what operations they performed. This can be done either with triggers or with the help of a new PL/SQL package, DBMS_FGA. |
|
| Index
Organised Tables: |
|
|
|
| Can now create and rebuild secondary indexes online. |
|
| Cursor
Sharing: |
|
|
|
|
By
setting a dynamic parameter, you can tell the system to share
SQL cursors for similar statements. This works by internally
replacing literals in the statements with bind variables,
thus automating what many application designers previously
had to do manually. |
|
|
|
 |
| Database
Resource Manager: |
|
|
|
| New functionality includes: the ability to create an active session pool for concurrent sessions; creation of an 'undo pool' to allocate specific amounts of undo to certain user groups; prevent execution of operations that will last longer than a predefined limit and automatic switching of users between groups according to limits set by the administrator. |
|
| Proxy
Authentication: |
|
|
|
| Enables you to authorise a middle-tier server to act on behalf of a client. |
|
| Multi-tier
Materalised Views: |
|
|
|
| You can now build updateable materialized views on top of other updateable materialized views. This provides the ability to create complex branched replication environments, e.g. for regional offices. |
|
| Cached
Execution Plans: |
|
|
|
| Oracle 9i stores execution plans for cached cursors in a set of new dynamic performance views. This means you can find out a particular execution plan at a particular moment when, for example, performance issues have been noticed. |
|
| Net8
and LDAP: |
|
|
|
| Net8 now supports an LDAP-Compliant directory service which acts as a central repository for database information, network components, user and corporate policies and preferences, user authentication and security information - thus replacing local client-side and server-side files. This essentially replaces ONAMES but with the added benefit of being able to support heterogeneous systems and networks. The process by which Net8 resolves an alias into a connect descriptor using LDAP is called "directory naming". |
|
| |
<<
previous page |
|
next
page >> |
|