 |
» |
|
|
|
There are many tools available to help you monitor
performance of HP-UX based servers, networks, and applications. Some
of these tools are included with HP-UX, some are downloadable from hp.com, and some are commercial products
from HP or other companies. This section discusses a few of these
tools. Tools for Monitoring the Performance of a Server | |
There are several applications you can use to
monitor the performance of an HP-UX based server: The following HP-UX commands can also help you
gather statistics about how the resources of your system are being
used: - iostat
iostat iteratively reports I/O
statistics for each active disk on the system. - sar
sar, the system activity reporter,
samples and reports on cumulative activity counters in the operating
system or from a previously recorded file. These values can give you
a rough idea of where the HP-UX is spending its time. - top
Supplied as part of HP-UX, top is
a utility that lists all of the processes currently running on a server,
ordered by their processing core usage. Those processes listed first
in the top output are consuming the most processing
time. top also shows global system load factors. - vmstat
The vmstat command reports certain
statistics kept about process, virtual memory, trap, and CPU activity.
It also can clear the accumulators in the kernel sum structure.
Tools for Monitoring the Performance of a Network | |
Monitoring the performance of a network can be
an involved process involving many different variables. For sophisticated
network troubleshooting and performance monitoring, HP offers the
OpenView Network Node Manager. For information about the Network Node
Manager’s features and capabilities, and how to acquire it
see: http://openview.hp.com/products/nnm/index.html If your needs are simply to verify communications
between two computers, you can use the ping command
which will send packets from one computer to another and time how
long it takes to receive a reply. You can do some basic tuning of
network performance by tweaking various network settings and running ping to see if the response time improves or worsens. For
example: Example 4-1 Testing network performance using ping To test a network connection between two systems called “thissystem” and “thatsystem”, from a local command prompt on “thissystem” enter the command: /usr/sbin/ping thatsystem PING thatsystem.xxx.yyy.com:
64 byte packets
64 bytes from 10.17.123.456: icmp_seq=0. time=1. ms
64 bytes from 10.17.123.456: icmp_seq=1. time=0. ms
64 bytes from 10.17.123.456: icmp_seq=2. time=0. ms
64 bytes from 10.17.123.456: icmp_seq=3. time=0. ms |
Ping will keep sending/receiving packets until you stop it by using
the interrupt character (usually CTRL-C). It will then terminate the
packet sending and report the final performance statistics: ----thatsystem.xxx.yyy.com PING Statistics----
4 packets transmitted, 4 packets received, 0% packet loss
round-trip (ms) min/avg/max = 0/0/1 |
Tools for Monitoring the Performance of Applications | |
Applications running on HP Integrity Servers can
be extensively profiled using the HP Caliper performance monitoring
tool (Caliper). While Caliper is capable of monitoring whole-server
performance, it is primarily an application profiling
tool. Caliper makes extensive use of hardware features
on HP Integrity Servers and can run on both HP-UX 11i and Linux operating
systems. |
Printable version
