HP Command View EVA Software - Manual (Page 93)

This creates the evapmfilt.conf file.
4. To filter data for specific virtual disks, enter the following command with the virtual disk names:
pfvd [vdisk]*
You can use friendly names in these commands.
5. When filtering configuration is complete, restart Windows Performance Monitor. When you
click Add counters in a new Windows Performance Monitor session, the evapmfilt.conf
file is detected and data is filtered accordingly. Until you turn filtering off, the filters you
configure remain active in successive sessions. (A message indicates that filtering is active.)
To make changes, follow the same procedure and use the same commands to update the virtual
disk or storage system names.
To stop filtering:
1.
Stop Windows Performance Monitor.
2. Open the command line interface.
3. Enter the following command:
pfd
4. Restart Windows Performance Monitor.
Objects and counters
Use the objects and counters in Windows Performance Monitor to gather performance metrics.
Objects are the items you can monitor, such as virtual disks, hosts, and controllers. The counters
characterize the workload and performance for each object.
The following identifying information is common to several objects:
•
Ctlr—The controller for which metrics are being reported. This field shows the last four digits
of the controller serial number.
•
Node—The storage system from which data has been collected.
•
GroupID—The disk group to which the virtual or physical disk belongs.
Not all metrics that are available in the command line interface are available in Windows
Performance Monitor.
HP EVA storage array
The HP EVA storage array object provides information about the total workload on the array.
The counters are:
•
Total Host Req/s —The total number (per second) of host-initiated read and write requests
to the controller pair.
•
Total Host MB/s —The total rate (per second) at which data is read from and written to
the disk by the controller pair.
HP EVA storage controller
The HP EVA storage controller object provides information about controller processor and host
data transfer utilizations.
The counters are:
•
% Processor Time —The percentage of time that the central processing unit on the controller
is active. A completely idle controller shows 0%. A controller saturated with activity shows
100%.
•
% Data Transfer Time —Similar to % Processor Time except that it does not include time
for internal processes not related to host-initiated data transfers. For example, it does not
include time for sparing, leveling, snapclones, snapshots, replication traffic, virtual disk
Using the graphical user interface 93