Silicon Graphics InfiniteStorage 4000 Series, Настройка и обслуживание

Страница: 186 / 276

174 SSD Cache Performance Modeling
The performance modeling tool provides an estimate of performance using these
metrics:

Cache hit percentage

Average response time
NOTE
Performance modeling does not survive a controller reboot.
Starting and Stopping SSD Cache Performance Modeling
To start a performance modeling operation, use this command:
start ssdCache [ ssdCacheName ] performanceModeling
Enclose the identifier in square brackets ([ ]). If the SSD cache name contains special
characters or consists only of numbers, you also must enclose the identifier in double
quotation marks (" ") inside square brackets.
After you start the performance modeling operation it will continue to run until you
stop the operation. To stop the performance modeling operation, use this command:
stop ssdCache [ ssdCacheName ] performanceModeling
This command immediately stops collecting the performance modeling data and
displays the data that has been collected. The data is displayed in the same window
from which you ran the command. If you prefer, you can save the performance
modeling data to a file by running this command:
stop ssdCache [ ssdCacheName ] performanceModeling
file=" fileName "
fileName
is the file path and the file name to which you want to save the
performance modeling data. Enclose the file name in double quotation marks (" "), as
shown in this example:
file="C:\Program Files\CLI\logs\performance.csv"
The default name of the file that contains the performance modeling data is
readLinkStatus.csv
. You can use any file name, but you must use the
.csv
extension.
The performance modeling metric data are not available until you stop the
performance modeling operation.
You can see the performance modeling data only after you have stopped the
performance modeling operation. You cannot get intermediate results while the
modeling is being run. When you stop the performance modeling tool, a graphical
representation of the results appears, but you can view the results in tabular format by
saving the data to a
.csv
file. To compare different results, you can run several
performance modeling operations and save each result to a
.csv
file for each result.