Introduction

Using crontab in UNIX is the ideal way to automate periodic NNM maintenance tasks. Many of these tasks are scheduled during the off hours to minimize operational impacts. Any output generated by these automated tasks can be logged or e-mailed to the NNM administrator. Some routine maintenance, such as map customization, remains as a manual task.

The scope of the daily backups can be as simple as running the ovbackup.ovpl script, which duplicates data in predefined directories. Entire volumes can also be archived to another system, and site-specific files and directories not in the usual NNM directories should be backed up.

Database repairs should be automated and executed at least weekly. These repairs are disruptive to NNM service and should be done during the off hours. The ovw -mapcount -ruvDR and ovtopofix -cshv commands can be automated with a crontab entry.

Moving newly discovered devices, subnets, and servers to their proper containers is a manual task for the map builder. Their icons appear in the new object holding area of the Internet submap. Map customization is typically done daily.

The NNM map builder backs up the customizations for the Internet submap after changes are made. This is important because the work effort to recreate this map from scratch is both daunting and time-consuming.

Updating MIBs for new devices introduced to the network may be automated, but the process is subject to MIB file syntax errors and should be executed manually to ensure success. New and updated MIBs may be loaded before a new device is installed to ensure that proper autodiscovery occurs.

Deleting unwanted maps is a task that may be automated for off hours execution. Well-run sites allow only properly named maps to exist. Keeping the map count down increases system performance and enhances its reliability.

Reviewing configuration alarms is not an NNM system maintenance task but a network maintenance task. This alarm category reports device configuration errors and represents a highly proactive task that should be executed frequently during the day.

Perusing log files and application alarms may be an automated task. Many of the problems NNM may encounter are logged, while others appear in the application alarm category of the event browser.

A review of MeasureWare data helps identify NNM systems in need of hardware upgrades. This is a task best done weekly using PerfView to examine resource utilization graphs.

Finally, you should peruse the ITO alarms as they occur to address any automated actions. As new automated actions are developed, they should be incorporated into ITO’s monitoring routine.