spotlight on oracle试用 (转)
spotlight on oracle试用
(1)安装
首先在PC上安装Oracle客户端。
我的选择是:Oracle9i Management and Integration 9.0.1.0.1
[描述是安装Management Server,管理工具,Oracle Internet Directory, oracle Integration Server, 网络服务,实用程序和基本的客户机软件]
再次选择:自定义
再次选择:安装Oracle Net Service 9.0.1.1.1中的Oracle Net Listener 9.0.1.1.1
然后结束安装
(2)配置
安装完成之后,有SQL Plus,有Network Configuration Assistant, Net Manager
可以通过Network Configuration Assistant配置[本地网络服务名
或者通过Net Manager的[本地]-[服务命名]添加[服务名],然后连接到Oracle数据库.
(3)测试
通过SQL Plus和第二步配置的服务名,连接到数据库服务器.
测试使用正常.
(4)安装Spotlight on oracle程序
(5)配置连接到Oracle数据库服务器
在主界面中选择[Connect Manager]
再次选择[New Connection]
启一个名字
在[connection string]处选择第二步配置的服务名,输入数据库的用户名和密码.
在[Monitor OS]处输入OS的IP地址,OS的用户名和密码.
(6)启动
在添加好的[连接]上选择[连接]
数据库的性能图表就显示了出来.
(7)性能参数
在性能的全体图上,非常形象的展示了数据库的整体处理过程的监控
客户端 服务进程(前台进程) (系统全局内存区) (后台进程) 物理存储
Sessions <–> Server Process <————————————————–> Storage
用户(活动用户) PGA SGA Backgroup Proecess
Dedicated <—-> Buffer Cache(Hit Ratio) <–> DBWR1 <—-> DB Files
Shared (Keep Pool,Recycle Pool)
Dispathers Redo Buffer <–> LGWR1 <—-> Redo Logs
Parallel Query Shared Pool(Uesed)
JOB Queue Java Pool
Large Pool ARCH1 <—-> Archive Log[
[通讯]
在客户端和服务前台进程之间
(1) < SQL*Net Send(KB/S): The rate at which data is being transferred from the server across the SQL*Net interface
(2) > SQL*Net Received(KB/S): The rate at which data is being transferred to the server, across the SQL*Net interface.
在服务前台进程和SGA之间
(3) < Logic reads/s (blks/s): The rate at which blocks are read from the SGA by all server processes.
(4) > Block changes/s (changes/s): The rate at which changes are made to blocks in the SGA by all server processes. The Lock Wait alarm becomes current on this component if updates are being blocked by locks.
(5) > Redo Buffer Entries/s (Entries/s): The rate of redo buffer entries made by all server processes.
(6) > Parse Requests/s (Requests/s): The rate of SQL parse requests per second by all server processes. The Parse ratio alarm becomes current if the ratio of parse requests to execute requests exceeds a threshold
(7) > SQL execution rate (Executations/s): The number of SQL execution requests per second by all server processes.
服务后台进程在SGA和物理存储之间
(8) > Physical Writes/s (blks/s): The rate at which modified blocks are written from the SGA to disk by the DBWR processes.
(9) > Redo Writes/s (blks/s): The rate at which redo log entries are written to the redo log files by the LGWR processes.
在物理存储和服务器前台进程之间
(10) < Physical Reads/s (blks/s): The rate at which blocks are read from disk by all server processes. The Datafile read time alarm becomes active if the average I/O time for these reads exceed a threshold.
[名词]
(1)Total PGA currently in use
The total amount of PGA (Program Global Area) currently in use by all active sessions.
For versions of oracle up to and including oracle 8i, it is sum of PGA used in all active sessions. From oracle 9i onwards, this value represents the "total PGA inuse" statistic in V$PGASTAT.
(2) Dedicated servers
Dedicated server processes that perform work on behalf of a single client process. The number of servers varies as users log in and out of the database
(3)Shared Servers
Shared or multi-threaded servers (MTS) perform work on behalf of more than one client process. The number of shared servers varies depending on load between the values of the configuration parameters MTS_SERVERS/SHARED_SERVERS and MTS_MAX_SERVERS/MAX_SHARED_SERVERS.
If a high proportion of MTS are busy, then the Multi-threaded server alarm becomes active. For advice on how to deal with this alarm, see the help topic Dealing with MTS contention.
(4) Dispatchers
MTS dispatchers coordinate the allocation of shared servers to client tasks.
The number of dispatchers varies depending on the load between the values of the configuration parameters MTS_DISPATCHERS/DISPATCHERS and MTS_MAX_DISPATCHERS/MAX_DISPATCHERS. If a high proportion of MTS dispatchers are busy, then an alarm may become current on this component. For more information about this alarm, see the help topic Dealing with MTS contention.
(5) Parallel Query
Parallel query servers support parallel execution of queries and (in oracle8) DML statements. The number of servers varies depending on load between the configuration parameters PARALLEL_MIN_SERVERS and PARALLEL_MAX_SERVERS.
(6) Job queue
Job queue server processes are responsible for running PL/SQL commands submitted to the oracle job queue via the DBMS_JOB package. The number of job queue processes is determined by the configuration parameter JOB_QUEUE_PROCESSES. If a high proportion of job queue servers become busy, then the Job processes busy alarm becomes current on this component.
(7) Buffer cache Hit Ratio
Buffer Cache Hit Ratio is the percentage of logical reads against total reads. It is the performance indication of the efficiency of the Buffer Cache.
A number of alarms may become current on this component. They include the Buffer Cache Hit Ratio alarm, Cache Buffer LRU Chains Latch alarm, Cache Buffer Chain Latch alarm and the Buffer Busy Wait alarm.
(8) Buffer cache
In order to avoid a disk I/O, the buffer cache is used to cache frequently-accessed data blocks that may subsequently be required.
In oracle versions before oracle9i, the size of the buffer cache is controlled by the parameter DB_BLOCK_BUFFERS. In later versions, the relevant parameter is DB_n K_CACHE_SIZE.
(9) Keep pool
The Keep Pool keeps objects inside Buffer Cache, without be aged out. The size of Keep Pool is controlled by parameter BUFFER_POOL_KEEP pre 9i and DB_KEEP_CACHE_SIZE since 9i.
(10) Recycle pool
The Recycle pool ages out objects inside it quickly. The size of Recycle pool is controlled by parameter BUFFER_POOL_RECYCLE in oracle versions before oracle 9i and DB_RECYCLE_CACHE_SZIE since 9i.
(11) Redo Buffer
The redo buffer contains redo entries that must eventually be written to the redo log.
Alarms can become current if processes spend time waiting for space in the redo buffer (Log buffer space wait alarm), or for redo buffer latches (Redo allocation and copy latch alarms).
(12) Shared Pool
The shared pool caches SQL statements, PL/SQL programs, object definitions, and session memory for MTS sessions.
A Spotlight user with Alter SYSTEM privileges can flush data manually from the shared pool. To do so, right-click on the Shared Pool or Shared Pool Used component and choose Flush from the shortcut menu. (This option exists only for oracle versions where manual flushing is supported.)
Note: Flushing the shared pool may adversely affect the performance of the database in the short term.
(13) Java pool
The Java pool caches class definition, Java methods, and Java objects. The size of the Java pool is controlled by the JAVA_POOL_SIZE paramater.
(14) Large pool
The Large pool shows the size of the large pool allocation heap. It is used in MTS for session memory. It can be used by parallel execution and backup processes. The size of the large pool is controlled by the LARGE_POOL_SIZE parameter.
(15) Database writer
The DBWR writes the buffer cache blocks from SGA to data files on hard disks.
The number of DBWR processes is determined by the parameter DB_WRITERS (Oracle7) or DB_WRITER_PROCESSES (Oracle8 and oracle8i). This process may display an alarm if it is determined that the DBWR process is causing delays to other processes. The possible alarms are the Free Buffer Waits alarm and the Write Complete Wait alarm.
(16) Log writer (LGWR)
The Log writer writes the redo buffer blocks from SGA to redo log files on the hard disk.
The number of LGWR processes is set at 1 in oracle 7. It may be configured in oracle 8 (but not oracle 8i) by using the LGWR_IO_SLAVES parameter. An alarm can become current on this process if log switch waits occur (Log Switch Time alarm), or if the average redo I/O time exceeds a threshold (Average Redo Write Time alarm).