HP 3000 Manuals HP 3000 Manuals
Chronological Detail
This section contains examples of the following chronological detail
report phases:
* PARM file global change notifications.
* PARM file application addition and deletion notifications.
* SCOPE OFF time notifications.
* Application-specific summary report.
* SCOPE-generated notes.
* USER-generated notes.
* ALARM events.
PARM File Global Change Notifications
You must specify DETAIL=ON and have GLOBAL data in the log file.
This report can be generated any time SCOPE is started or is instructed
to resample its PARM file. If the current PARM file collection
parameters differ from the parameters when SCOPE ran last, a global
change notification can occur.
The following is an example of the change notifications that occur when
four new disc drives are added to the system.
01/26/90 16:43 THE NUMBER OF DISC DRIVES CHANGED FROM 8 TO 12
01/26/90 16:43 DISC # 9 WAS ADDED AS LDEV 18
01/26/90 16:43 DISC # 10 WAS ADDED AS LDEV 30
01/26/90 16:43 DISC # 11 WAS ADDED AS LDEV 31
01/26/90 16:43 DISC # 12 WAS ADDED AS LDEV 32
PARM File Application Addition/Deletion Notifications
You must specify DETAIL=ON and have APPLICATION data in the log file.
User-defined applications can be added or deleted each time SCOPE is
started or is instructed to resample its PARM file. If you find an
application name that does not match the last set of applications, you
can list an application addition, deletion, or change notification. If
the name of an application has not changed from that previously reported,
it is not listed again.
NOTE Application definitions are not checked for changes at this time.
They are listed when an application name is changed, but any change
to an existing application's definition without an accompanying
name change is not detected.
01/31/90 21:11 APPLICATION 4 "COMPILES " WAS ADDED
FILE=COBOL@.PUB.SYS; FORTRAN.PUB.SYS; FTN@.PUB.SYS; SPL
FILE=PASCAL.PUB.SYS
This example indicates a new application was started.
SCOPE OFF Time Notifications
You must specify DETAIL=ON.
If an extracted files contains only summary information, times are
rounded to the nearest hour.
01/29/90 11:00 - 01/29/90 12:34 COLLECTOR OFF ( 01:34:04)
The first date and time (01/29/91 11:00) indicate the last valid data
record in the log file before SCOPE was restarted. The second date and
time (01/29/91 12:34) indicate when SCOPE was restarted.
The last field (in parentheses) shows how long SCOPE was not running.
The format is ddd/hh:mm:ss, where ddd are days, and hh:mm:ss are hours,
minutes, and seconds. Zeros to the left are deleted.
In this example, SCOPE was off on January 29, 1991 between 11:00 AM and
12:34 PM. The summary information shows that data was not collected for 1
hour, 34 minutes, and 4 seconds.
Application-Specific Summary Report
You must specify DETAIL=ON and have APPLICATION data in the log file.
This report can help you define applications. Use the report to identify
applications that are accumulating either too many or too few system
resources and those that could be consolidated with other applications.
Applications that accumulate too many system resources might benefit by
being split into smaller pieces.
You should define applications in a way that helps you make decisions
about system performance tuning. It is unlikely that system resources
would accumulate evenly across applications.
The application-specific summary report is generated whenever the
application definitions change to allow you to access the functionality
of the applications before and after the change.
A final report is generated for all applications. This report covers
only the time since the last report and not the entire time covered by
the log file.
PERCENT OF TOTAL
APPLICATION RECORDS CPU DISC TRANS
-------------------- ---------- ------ ------ ------
OTHER 22385 45.7% 20.9% 63.0%
Resource Sharing 7531 6.0% 2.2% 17.1%
SPOOLING 13813 2.4% 0.3% 0.0%
ON-LINE COMPILES 13119 2.9% 1.7% 0.1%
BATCH COMPILES 8429 2.9% 0.1% 2.2%
ORDER ENTRY 387 0.1% 0.0% 0.0%
ELECTRONIC MAIL 6251 3.8% 1.3% 9.6%
PROGRAM DEVELOPMENT 3141 9.1% 2.4% 0.6%
RESEARCH DEPARTMENT 3968 8.7% 2.0% 6.0%
BILL OF MATERIALS 336 0.6% 0.2% 0.1%
FINANCIALS 1080 5.0% 1.5% 0.5%
MARKETING DEPT 2712 12.9% 67.3% 0.0%
GAMES 103 0.1% 0.0% 0.6%
-------------------- ---------- ------ ------ ------
ALL USER APPLICATIONS 73.1% 54.3% 79.1% 37.0%
SCOPE-Generated Notes (Collector Shutdown)
You must specify NOTES=ON (or NOTES=102) and have GLOBAL data in the log
file.
01/30/90 10:56 NOTE [102] SHUTDOWN REQUESTED BY PURGING "RUN" FILE
SCOPE can enter a note record in the global log file. Typically, it does
this to indicate an important event such as its voluntary termination.
In the example above, SCOPE notes that it is terminating voluntarily
because the RUN file was purged.
The number in brackets ([ ]) indicates the origin of the note and can be
filtered from this listing using the LEVEL option of the NOTES command.
User-Generated Notes
A note can be entered in the global log file by running the UTILITY
program and issuing the SCOPE NOTE command. These notes are displayed
during a SCAN of the log file if the NOTES=ON or NOTES=100 command is in
effect.
04/01/90 08:05 NOTE [100] SYSTEM MELT DOWN IN PROGRESS
Performance Alarm Events
Before you can see any performance alarm events, you must define alarms
in a file that can be accessed using the PARMFILE command. See chapter 7
for more information on defining alarms. The log files do not contain
these events, but the events are generated based on log file data
following the rules in the PARM file. If you use a different set of
alarm definitions during a SCAN, you will obtain a different set of alarm
events.
If alarms are defined and you have ALARMS=ON, an alarm-begin event will
be listed every time an alarm has met all its conditions for the
specified duration. When these conditions are no longer satisfied, an
alarm-end event will be listed. If an alarm condition is satisfied for a
period long enough to generate another alarm without having first ended,
a repeat event will be listed.
Each event listed will show the alarm number, how long the alarm has been
active, and the alarm ID. For example:
11/21/91 21:30 ALARM[ 4] BEGIN AFTER 10 MIN "MEMORY THRASHING"
11/21/91 21:40 ALARM[ 4] END AFTER 20 MIN "MEMORY THRASHING"
12/05/91 21:20 ALARM[ 5] BEGIN AFTER 10 MIN "CPU OVERLOAD"
12/05/91 21:30 ALARM[ 5] REPEAT AFTER 20 MIN "CPU OVERLOAD"
12/05/91 21:35 ALARM[ 5] END AFTER 25 MIN "CPU OVERLOAD"
If you specified ALARMS=ON and DETAIL=ON, you will see the actions (THEN
and FINISH) in addition to the alarm events. The actions will not be
performed, but they will be listed with any requested parameter
substitutions in place.
11/21/91 21:30 ALARM[ 4] BEGIN AFTER 10 MIN "MEMORY THRASHING"
:TELLOP Severe Memory Shortage (Swapping at 25.8/second)
11/21/91 21:40 ALARM[ 4] END AFTER 20 MIN "MEMORY THRASHING"
:TELLOP OK, relax, swapping is down to 2.3/second now.