disk/HUNTER addresses the common problem of “disappearing” disk space. Modern, complex AS/400 and iSeries applications can consume disk storage so rapidly and excessively that you see a “spike” in disk utilization. Finding which jobs are using up disk storage and for what objects, is a difficult and time consuming task - a task that, if not completed quickly enough, can lead to application slowdowns or system crashes.

Many application and database problems can lead to disk spikes, including:

1. Database queries that create large, temporary indexes (access paths)
2. Database queries that create large temporary spaces to implement hash joins or grouping
3. Applications that use heap storage but do not free it (memory leaks)
4. Java-based applications that create a large number of objects but that do not get reclaimed because of dangling object references
5. Applications that use the Integrated File System (IFS) and create large stream files but do not always delete them when they are done
6. Users who create copies of data in IFS as backups not realizing how much disk space they consume
7. Jobs caught in an infinite loop writing printing and creating very large spool files
8. Applications that write to log files that do not get cleared often or at all
9. Data collected by IBM performance and trace tools

disk/HUNTER provides real-time disk space monitoring, notification, analysis, and diagnostic reporting. This allows the system operator or application management team to react quickly when usage problems surface.

The disk/HUNTER utility’s main benefit is that it allows system administrators to react immediately when a disk spike occurs rather than having to go through a long, tedious, and error prone manual process which, if not done in time, could allow disk spikes to reduce system performance and availability.

IBM’s Management Central provides some basic monitoring and notification related to disk space usage. disk/HUNTER compliments the functions in Management Central by providing additional detail on the source of the disk consumption. Since Management Central allows users to define trigger actions when system events occur, you can configure disk/HUNTER to be invoked when defined thresholds are reached. Alternatively, you can use disk/HUNTER’s built-in monitor to achieve the same objectives.

disk/HUNTER monitors your system at all times, keeping an eye on disk space utilization. If the amount of disk space suddenly grows, disk/HUNTER reacts automatically - to track down information about the activity causing the problem. It then generates reports telling you what users, what jobs, and what objects are consuming the space.

All of this investigation is done in a matter of minutes, allowing your team to react quickly and with pinpoint accuracy. This allows you to avoid lengthy or dangerous problems while allowing focused actions to be taken that impact the fewest people.

The disk/HUNTER utility has been designed for minimal impact (less than 1% CPU) on your system while actively monitoring disk space use. If a sudden spike occurs, additional data collection and analysis is done to determine the root cause of the problem. This requires additional system resources, but should not take more than 10% of a single processor for more than 5 minutes on average. The specific amount of resource will depend on how you configure disk/HUNTER.

Currently disk/HUNTER works on V4R5 and higher.

The installation and configuration of disk/HUNTER is very simple. Typically, it takes less than half an hour from the time the CD is put in your iSeries or AS/400 to when disk/HUNTER is fully operational and monitoring your system.

You may specify how frequently disk/HUNTER checks on your disk storage. You may also specify the threshold at which disk/HUNTER will notify you of a problem. This threshold can be set either in actual megabytes consumed, or as percent of total disk storage. For disk/HUNTER notification purposes, you may specify the message queue to receive a threshold notification message and an output queue for the diagnostic reports.

No. The change to disk/HUNTER is automatic and does not require the program to be ended and restarted.

Yes. The parameters can be changed on the start disk/HUNTER command. The parameters will be in effect until disk/HUNTER is ended or the configuration is run.

Yes. If a stream file in IFS grows so much that it triggers disk/HUNTER diagnostics and analysis, that file’s growth will appear in the object growth report.

disk/HUNTER is installed at many iSeries shops that need to be proactive about disk space management.

disk/HUNTER is part of a growing suite of utilities called insure/TOOLS. We also have integrated disk/HUNTER into our insure/MONITOR product as another real-time tool to help iSeries customers actively manage their systems and applications. Our first integrated TOOLS/MONITOR offering will be our high value insure/MONITOR for JDE OneWorld module which tracks and reports JDE OneWorld user profiles and includes the disk/HUNTER utility.

Yes. disk/HUNTER issues messages whenever disk free space has been reduced below the user-defined threshold. These messages have unique IDs, as well as substitution variables. Hence, they can easily be detected and processed by these applications.

Messages include notification of the triggered event, as well as the results of the system check performed by disk/HUNTER. The triggered event notification message (XDH0005) has severity level 99, which can cause this message to break at the operator's console whenever it is sent.