HP OpenView System Administration Handbook [Electronic resources] : Network Node Manager, Customer Views, Service Information Portal, HP OpenView Operations نسخه متنی

16.3 OTHER PERFORMANCE TOOLS

We have seen two of the most popular tools that collect performance data within the OpenView domain. This section describes some of the other tools that are available to monitor, collect, and analyze collected performance information. Some of the tools are built into the operating system free of charge. Others like Glance and the OV Performance Manager require a license to use.

16.3.1 OV GlancePlus Pack

GlancePlus is a real-time diagnostic tool that displays current performance data directly to a user terminal or workstation. It is designed to help identify and troubleshoot system performance problems as they occur. Figure 16-8 shows the contents of the Glance Application Group, which makes it easy to start, stop, configure and check the status of Glance. The data is displayed for only one node at a time.

Glance includes two user interfaces: one is character/text-based and the other is a Motif-based graphical interface. Both interfaces allow the user to customize the views of the metric data. Glance is bundled as a separate product or in a combination bundle with OVPA called the GlancePlus Pack. Glance is supported on the following operating systems: HP-UX, LINUX, Solaris, AIX, and Sinix.

Refer to the support documentation available at: http://support.openview.hp.com for the most current platform support information.

16.3.2 OpenView Performance Manager (OVPM) 4.x

OVPM 4.x replaces the product called PerfView and provides integrated performance management for multi-vendor distributed networks. OVPM runs on UNIX and Windows platforms, clients connect to OVPM via a web-based interface. It gives you a single interface and a common method for centrally monitoring, analyzing, and graphically forecasting resource measurement data supplied by OVPA.

OVPM also displays information from OV Service Reporter, OV Internet Services and OV Operations. There are three OVPM components as described here:

• Analyzer
  included with the base product, performs analysis of historical performance data for trend analysis.

• Monitor
  (add-on) receives, displays and filters alarms

• Planner
  (add-on) provides the means to forecast future performance in order to facilitate performance management.

16.3.3 Glance versus OV Performance Manager

Although Glance and OVPM both provide metric data in user friendly graphical format, the two programs have unique differences. The features of the two programs are compared in Table 16-2.

Table 16-2. Comparison of Glance and Performance Manager

If you need a more real-time view of the metric data, Glance would provide the best solution. If you plan to analyze the data you are collecting over a period of time to make investment, upgrade, migration, or configuration decisions, Performance Manager provides the best solution.

16.3.3.1 Terms and Definitions Used with Glance

Application Response Measurement (ARM)Run-time library APIs linked within an application program. During run-time the application transactions are collected and logged and available for graphing.

Transaction Tracking (TT)Collect, analyze and graph transaction response times, transaction counts, and service level violations that occur within an application during run-time. Requires a LOGTRAN entry in the OVPA parameter file, registration daemon (ttd), configuration file (ttd.conf), and customized application with ARM APIs.

16.3.4 OVPA Applications Integration with OVO

The OVO application bank has an application group called performance. Figure 16-11 shows the performance agent application group that contains (OVPA) applications that provide out-of-the-box integration with OVO. Some of OVPA application group components are listed here for reference:

• Check alarmdef

• Check parm

• Config alarmdef

• Config parm

• Config perflbd.rc

• Config ttd.conf

• List processes

• List versions

• Restart perf agent

• Start perf agent

• Start extract

• Stop perf agent

16.3.5 Network Node Manager

NNM has built-in performance tools that allow you to check the health of the managed devices in the domain. The tools are built into the menu of the root window of the OVO Administrators console. The information available from within NNM provides the following:

Collect the Performance Data
using snmpCollect

Configure Performance Monitoring Threholds
using snmpCol.conf

Performance Data in text format
using snmpColDump

Performance Data in graphical format
using xnmgraph

Remote Monitoring
using the MIB (RMON)

Capacity Planning
using the current topological information

Store and Report the Performance Data-using the Data Warehouse and Reporting tools

16.3.6 Databases

Many of the performance measures for the database environment within the OV domain can be captured with the OpenView Database SPIs. Refer to the specific database documentation for performance tools that are available.

16.3.7 Operating System Built-in Performance Tools

Commands built into the UNIX operating system can provide valuable information to help collect and analyze performance data for specific system resources. The programs outlined here do not represent every possible tool for all supported OpenView server and agent operating systems. Consult the specific system documentation (man pages on HP) for a detailed listing of the commands and command line options.

16.3.7.1 top

Displays and updates information about the top processes on the system. Summarizes the general state of the system (load average), quantifies amount of memory in use and free, and reports on individual processes active on the system.

top updates its displays at intervals. On multi-processing systems, top reports on the state of each CPU.

16.3.7.2 sar

Reports on cumulative system activity, including CPU utilization, buffer activity, transfer of data to and from devices, terminal activity, number of specific system calls used, amount of swapping and switching activity, queue lengths, and other kernel tables.

16.3.7.3 vmstat

Quantifies the use of virtual memory by processes on the system; also reports on traps and CPU activity.

vmstat produces CPU and virtual memory metrics on an interval specified at the command line.

16.3.7.4 iostat

Reports I/O statistics for active disks, terminals, and processor(s)

16.3.7.5 nice

The

nice command executes a command at a non-default CPU scheduling priority. A nice value influences a program's process scheduling priority. When the nice value is higher than the default schedule priority (19 on HP-UX), the schedule priority for the program is lowered. The nice value is used to compute the priority value; it does not set the priority value.

16.3.7.6 ps

Report the current process status. Use this command to determine if all of the OpenView, user, or other application processes are running. Here is an example of the

ps command specifically looking for the running Oracle processes:

ps ef|grep ora

16.3.7.7 size

Prints the total size of the object file, including the text, data, and unintialized data.

16.3.7.8 swapinfo t

The

swapinfo command is used to review the current configuration for device and filesystem swap. Sometimes when you run the OV system check report (more on this in Chapter 17) it will print a warning that there is insufficient swap space configured. System swap space is configured when the operating system is installed. This determines how much virtual memory is available when it becomes necessary for the process vhand to select pages in memory that can be temporarily paged out to the swap space on disk. The swap area also functions as a dump device when necessary.

16.3.7.9 SAM

SAM is the system administration tool provided with the HP-UX operating system. This tool is invaluable when you need quick access to the operating system tools to perform administration tasks. For example, setting up new volume groups, lan cards, users, and tuning the kernel. Activities performed by SAM are logged in the /var/adm/sam/sam.log

16.3.7.10 cron

Cron executes commands and programs at timed intervals. It is helpful for collecting data and performing other important tasks routine tasks regularly.

16.3.7.11 crontab l

List cron's scheduled jobs.

16.3.7.12 uptime/ruptime

Reports the local or remote system's load average.

16.3.7.13 time and timex

The

time command reports on the time it takes to execute a command. The

timex command also reports the time it takes for a command to complete along with additional details if invoked with the s or p options.

16.3.7.14 ipcs

Ipcs provides a print to standard out on semaphores message queues and shared memory. You may recall that some of the kernel parameters for OpenView are for semaphores.

16.3.7.15 bdf and df

The

bdf and

df commands show information on the current mounted file systems, total capacity, amount used, amount available, percent used, and percent available. These commands also display inode percent used, and the number of used and free inodes.