EMS Manual

Generating Standard Events
EMS Manual—426909-005
10-9
Task 4. Define Private Event Types for Your
Subsystem
Task 4. Define Private Event Types for Your
Subsystem
To use EMS events to support management functions not currently defined in
Section 9, Standard Events
, use this task to determine whether EMS events should be
used and, if so, the content of these events. This task ensures some consistency
between events you defined and standard events.
If you define management functions that could benefit others, forward your results from
this analysis to HP for possible inclusion in the standard events.
For help converting your existing events, see Task 5. Migrate Existing Events
on
page 10-13. You are not required to define the management functions for your existing
Usage
Threshold
Use the EMS_CHECK_USAGE_THRESHOLD_ procedure to determine if
this type should be sent for each monitored usage:
1. Identify the object and kind of usage to monitor in each event. Consider
all objects when selecting resources to monitor.
2. Provide current usage level, configured high level, and configured low
level information. The procedure maintains and supplies the rest of the
event information. Let the operator configure high and low usage levels
and the unit of measurement of each resource.
Operator
Attention
Needed/
Intervention
Completed
If you generate an Operator Attention Needed event, you must also generate
an Operator Intervention Completed event after the operator has performed
the function (or after you decided to cancel the function).
1. Identify the object (such as \nodea.$tape) and the kind of operation (such
as mount tape “1234”). The identified object is the device that needs
attention, not the affected object in your subsystem.
2. For an Operator Attention Needed event, create a new 16-bit ID
(ACTION-ID token) for each new request.
3. Use this ID for the Operator Intervention Completed event.
4. If your subsystem cannot automatically detect when an action is
complete, you must let the operator notify you so you can generate an
event to remove the outstanding request.
Transient
Fault
Transient faults are recoverable errors, such as a controller error, network
reset, memory error, and so on:
1. Define the transient fault types (for the ZEMS-TKN-TXFAULT-TYPE) your
subsystem reports.
2. Determine how often to report these errors in events.
3. Define token(s) describing the error, affected object, and diagnostic
information, if known.
For event
type... Customize as follows: (page2of2)