Use the Summary page to monitor types of I/O activity on the database and view I/O behavior that may cause problems in database performance.
To open the Summary page
Select the Spotlight on Oracle RAC connection in the Spotlight Browser.
Click I/O | Summary.
Category?
Use Category to choose how to view I/O activity on the monitored database.
Category | Description |
---|---|
File Type |
File Type groups I/O activity by the files on which I/O operations are performed. |
Wait Type |
The default Wait Type category groups I/O according to the types of I/O performed. |
Throughput - requests Chart
Throughput—Requests charts the rate of I/O requests according to the category selected. As the number of requests approaches the limit of the I/O system being used for the database, latency will start to degrade.
Values that are consistently high over the displayed period are those most likely to contribute to performance bottlenecks.
Category | Description |
---|---|
File Type |
(Oracle 11g and later) The data series in the chart are derived from V$IOSTAT_FILE and show the rates of I/O requests for types of files in the database. (Oracle 10g) The data series are derived from V$FILESTAT and V$TEMPSTAT. |
Wait Type |
The default Wait Type category groups I/O requests according to the types of I/O performed on the database. Spotlight generates the data series in this chart from information in V$SYSTEM_EVENT. (See also Activity | Summary Page | Event Waits Chart.) |
Throughput - transfer rate (by File Type) Chart
For the category selected, Throughput—Transfer Rate shows (in MB/s) the rates at which data moves through the I/O subsystem.
View the data in this chart and compare each data series to the corresponding series in Throughput—Requests. (A high number of requests, but a low transfer rate could indicate a lot of single block reads. If the transfer rate is approaching the limit of the I/O system, this could cause a performance bottleneck.)
Notes:
Latency Chart
For the category selected, Latency shows the average time delay experienced while the database performs different types of I/O activities.
Values that are consistently high over the displayed period are those most likely to contribute to performance bottlenecks.
To maximize response time to end users, the latency should remain low – in general a latency below 10ms will give users good response time. As the latency increases, the response time may degrade. This time will vary from database to database – for example in a large batch job – the latency might not be as important as the throughput, but when there is a lot of users on the system, a low latency is important to ensure response time.
Total Service Time Chart
For the category you have selected, Total Service Time shows (in milliseconds per second) the rate at which the database is servicing different types of I/O activities.
The chart indicates the overall I/O workload on the database, and a consistently high value in any series may indicate an I/O bottleneck.
You can also compare the data in this chart to the corresponding data in Latency. (High latency for a type of I/O activity may not be significant if the total service time is small.)
If the service time for I/O is a major proportion of the DB time – there may be an I/O performance problem, especially if this is combined with a high latency. If the DB time is a lot higher than the I/O service time, and the latency is low, then I/O is probably not causing a performance problem.
Notes:
© 2024 Quest Software Inc. ALL RIGHTS RESERVED. Termini di utilizzo Privacy Cookie Preference Center