HP Matrix Operating Environment Software - Manual (Page 18)

1 Data is obtained from some managed system types using more than one collector. This is done where the data from
both collectors provide a broader or more accurate view than can be obtained from using only one collector on that
managed system type. For example, where both natively collected data (agentless) and data from the Integrity Virtual
Machines Provider are used, the HP VM agent provides info about the VM that allows Capacity Advisor to obtain
more accurate CPU data for the VM, and the agentless method provides metrics other than CPU utilization.
NOTE: The supported collection configurations for a managed node are normally set up by
the Insight managed system setup wizard. If the capcollect command is run before the wizard
is run on a managed node, and agentless data collection is supported for the managed node, the
node will be configured for agentless collection at that time. Because no agentless data is yet
available, however, the message:
Error: At the time data collection was run, system "system-name" was not
configured for agentless data collection. It has been configured and data
collection will begin shortly. No utilization data is available now.
Try collecting again in 5 minutes.
will appear in the command output. Subsequent collections will retrieve collected data beginning
at the time the capcollect command was run on the managed node.
Comparison of agentless and UP data collection
Table 2-2 (page 18) compares the operational characteristics of agentless data collection and that
of the Utilization Provider (UP) when used to collect data from systems licensed for Capacity
Advisor.
• Agentless Data Collection . This method leverages native data collection already occurring
on the managed server(s).
• WBEM Utilization Provider . This method uses an agent to gather the data and provide it
to the CMS.
Table 2-2 A comparison of agentless data collection and use of Utilization WBEM Provider
Utilization Provider (UP) Agentless data collection
Equally useful in data quality and accuracy.
Available for HP-UX and OpenVMS systems Available for Microsoft Windows, Linux, and non-HP
systems
Provides the basic set of resource utilization metrics for
CPU, memory, network I/O, and disk I/O. Provides the basic set of utilization metrics for CPU,
memory, network I/O, and disk I/O, plus additional
metrics for reporting specific activities in the operating
system, disk, and network: CPU queue length, page
faults/second, disk I/Os per second, LAN packets/second,
number of disks, disk space total, disk space used.
Requires hosting and updating the UP agent Leverages existing data on the system; no additional agent
hosting or update required
Not susceptible to downtime in the network or the CMS.
Collected data is held for a maximum of 30 days by the
UP for transfer to the database on the CMS Susceptible to network or CMS downtime as data is
continuously collected by the CMS across the network
from each managed system
Slight additional load on CMS CPU or memory resources
incurred once per day Small additional load on the CMS CPU and memory
resources incurred by the ongoing operation of the
agentless data collection service
Slight additional load on managed system resources, due
to local queries from the UP agent every 5 minutes, and
storage/retrieval of historical metrics on the local file
system Slight additional load on managed system resources due
to remote queries from the collector every 5 minutes
18 Features