Siemens U3585-J-Z125-8-76 1, SM2 User Manual
1 Preface
1.1 Brief description of the SM2 monitoring system
The SM2 monitoring system provides the user with statistics on the performance of the DP
system and on resource utilization.
It consists of the following components:
● the SM2 monitoring program
● the SM2U1 utility routine
● the SM2R1 analysis routine
The SM2 monitoring program collects statistical data and outputs it either to a terminal and/
or to a file. The first option is ideally suited for real-time (online) monitoring. If the data is
output to file, the stored data can be analyzed (offline) at a later time.
The SM2U1 utility routine manages the SM2 output files and is occasionally used as a file
conversion leader for the SM2R1 analysis routine.
The SM2R1 utility routine manages the statistical data registered by SM2 and written to a
system-global SM2 output file.
The following products are also available and can be used to edit and display SM2
monitored data:
● The SM2R1-PC analysis routine is used to graphically display the monitored data on a
PC.
● The SM2ONLINE-PC monitoring program is used to simultaneously monitor and
display monitored data from several BS2000/OSD computers on a PC.
● The SM2-PA analysis routine is used for user-specific SM2 output files.
U3585-J-Z125-8-76 1
Summary of contents Preface
1.2 Target group
This manual is intended for system users who need information on the current system
workload. It will be of particular interest to computer center and systems support staff who
wish to assess the performance of their configuration and isolate bottleneck areas by
means of long-term monitoring.
1.3 Summary of contents
This is Volume 1 of a set of two manuals and describes the SM2 monitoring program.
Volume 2 describes the SM2U1 conversion routine and the SM2R1 analysis routine, and
gives an overview of the other analysis routines. These routines allow you to carry out
performance analyses in BS2000.
Volume 1: "Administration and Operation"
Chapter 1, Preface,
specifies the target group of the manual and provides pointers on how to use it.
Chapter 2, The SM2 monitoring program,
describes the monitoring program and gives a overview of the terminology used. This
chapter provides an introduction to system monitoring and optimization, and provides
suggestions on how to carry out performance monitoring with SM2.
Chapter 3, SM2 monitoring programs,
describes the SM2 monitoring programs and measurement data.
Chapter 4, SM2 operation,
describes SM2 operation and statements for real-time and background monitoring. The
statements are presented in two sections: an administration section and an ev aluation
section. The first section describes the statements for the SM2 administrator for
controlling the monitoring process and outputting the monitored data to the SM2 output
file. The second section describes the statements availab le to nonprivileged SM2 users
for selecting and controlling reports.
Chapter 5, SM2 screen outputs,
describes the individual output forms (SM2 reports) for monitored data and the SM2
information screens.
Chapter 6, Notes on SM2 operation,
contains important information to be taken into account when using the monitoring
system.
2 U3585-J-Z125-8-76
Preface Summary of contents
Chapter 7, Variables reports,
contains a table detailing variables and report groups.
Chapter 8, Messages,
lists all the program messages, together with an explanation and the action to be tak en
in each case.
Chapter 9, Installation,
describes how to install SM2.
Chapter 10, Appendix,
contains descriptions of the SDF syntax and overviews of the SM2 statements.
At the back of the manual, you will find a glossary, lists of abbre viations, figures and tab les,
and related publications, and an index.
Volume 2: "Analysis and Display of SM2 Monitored Data"
Chapter 1, Preface,
specifies the target group of the manual and provides pointers on how to use it.
Chapter 2, SM2U1 utility routine,
describes the functions and operation of the SM2U1 conversion routine.
Chapter 3, SM2R1 analysis routine,
describes the functions and operation of the SM2R1 analysis routine and the records
of the SM2R1 transfer file.
Chapter 4, Other analysis routines,
provides a short description of other analysis routines for SM2 monitored data which
may be ordered.
Chapter 5, Variables reports,
contains tables detailing variables, report groups, report names and reports on
monitored variables (with information on the variables in question).
Chapter 6, SM2 program interfaces,
presents the C interfaces SM2GMS and SM2GDAT and the assembly language
interface PFMON.
Chapter 7, Appendix,
contains descriptions of the SDF syntax and the data structures for C.
At the back of the manual, you will find a glossary, lists of abbre viations, figures and tab les,
and related publications, and an index.
U3585-J-Z125-8-76 3
Changes made since the last version Preface
1.4 Changes made since the last version of the manual
SM2 V ersion 13.0A (J une 1999) incorporates the follo wing major changes as compared to
the previous version (V12.0A, December 1996):
SM2
● New monitoring programs
– BCAM-CONNECTION
This monitoring program is used to record monitored data for defined connections.
– DISK-FILE
This monitoring program is used to record inputs and outputs for selected disks.
– HSMS
This monitoring program is used to record monitored data on the migration and
retrieval of files.
● Changes to the monitoring programs
– CHANNEL-IO
Monitored data is also output in the CHANNEL report and at the SM2GDAT
interface.
– DISK
The number of possible devices has been increased to 256.
– FILE
File-specific access times are also output.
–PFA
Monitored data from SYMMETRIX controllers that do not comply with the PFA
concept is also recorded.
– SERVICETIME
The number of possible devices has been increased to 256.
– TASK
The number of devices that can be monitored has been increased to 64.
4 U3585-J-Z125-8-76
Preface Changes made since the last version
● New reports
– BCAM CONNECTION
Outputs information on data that has been received/sent and on bucket-specific
times.
– DISK FILE
Outputs files showing the inputs and outputs per second for the monitored disk
devices.
– SYMMETRIX CONTROLLER
Outputs monitored data on the use of SYMMETRIX controllers.
– SYMMETRIX DEVICE
Outputs monitored data on the use of the individual devices of a SYMMETRIX
controller.
● Modified report output for the CHANNEL, DAB, DAB CACHE, FILE, GS, ISAM, PFA
CONTROLLER and VM2000 reports
● New administrator statements
– ADD-BCAM-CONNECTION-SET
Specifies the connection set for the new BCAM-CONNECTION monitoring
program.
– REMOVE-BCAM-CONNECTION-SET
Specifies the connection set to be excluded from the monitoring process for the new
BCAM-CONNECTION monitoring program.
– SET-BCAM-CONNECTION-PARAMETER
Specifies the global monitoring parameters for the BCAM-CONNECTION
monitoring program.
– SET-DISK-FILE-PARAMETER
Defines the disk devices which are to be monitored by the new DISK-FILE
monitoring program.
● Modified and enhanced administrator statements
– SET-CHANNEL-IO-PARAMETER
Three-digit hexadecimal numbers can be specified for the CHANNELS operand.
– SET-DISK-PARAMETER
The DEVICES operand has been extended to accept 256 device names.
U3585-J-Z125-8-76 5
Changes made since the last version Preface
– SET/MODIFY-COSMOS-PARAMETER
The new MEASUREMENT-TIME operand can be used to specify the duration of the
monitoring process.
In the TSN operand, up to 16 TSNs can be specified.
In the TSN, USER-ID, JOB-NAME and CATEGORY operands, the last character
may be an asterisk (*).
– SET-SERVICETIME-PARAMETER
The DEVICES operand has been extended to accept 256 device names.
– SET-TASK-PARAMETER
The DEVICES operand has been extended to accept 64 device names.
– SHOW-ACTIVE/DEFINED-PARAMETER
The TYPE operand has been extended to include the new monitoring programs
BCAM-CONNECTION and DISK-FILE.
– START/STOP/CHANGE-MEASUREMENT-PROGRAM
The TYPE operand has been extended to include the new monitoring programs
BCAM-CONNECTION, DISK-FILE and HSMS.
● New, enhanced and obsolete user statements
– The SELECT-SYMMETRIX-CONTROLLER user statement (select the
SYMMETRIX controller for the SYMMETRIX-DEVICE report) has been included in
the manual for the first time.
– The REPORT user statement has been extended to include the values BCAM,
DISK_FILE (DFILE) and SYMMETRIX (SYM).
– The CHANNEL user statement is now obsolete, since the CHANNEL report can
now output follow-up screens.
● SM2 program interface
In the SM2GDAT macro, the variable ’unsigned long buffer_flags’ is now incompatible.
The BCAM-CONNCETION, CHANNEL-IO, DISK-FILE and HSMS data buff ers are now
available .
● Changes to the SM2 output file
In SM2 output files with freely selectable file characteristics, the block length must be
16.
● Dynamic IO reconfiguration
SM2 recognizes a dynamic IO reconfiguration and, if necessary , automatically changes
the number of objects being monitored. This is described in a new section in chapter 2.
● The constraints on SR2000 have been incorporated. For further information, please
refer to section 5.8 of the manual "OSD-SVP V2.0 User Interfaces on Systems with
RISC Architecture" [20].
6 U3585-J-Z125-8-76
Preface Notational conventions
SM2U1
● Extended statement: SELECT-MEASUREMENT-GROUPS
The statement has been extended to include
– the new monitoring programs BCAM-CONNECTION, DISK-FILE and HSMS
– the new report groups SYMMETRIX-CONTROLLER and SYMMETRIX-DEVICE.
SM2R1
● New statement: PRINT-HSMS-STATISTICS
Data from the HSMS monitoring program is displayed in the form of tables.
● Enhanced statement: PRINT-REPORTS
The statement has been extended to include the report groups BCAM-CONNECTION,
SYMMETRIX-CONTROLLER and SYMMETRIX-DEVICE.
The report groups CPU, DAB, DEVICE and FILE have been extended.
The operand REPORT-LIST=CHANNEL has been modified to take account of the
maximum channel number of 511.
● Extended statement: PRINT-DISK-STATISTICS
The DEVICES operand has been extended to accept 256 device names.
● Modified statement: PRINT-TASK-STATISTICS
The output has been modified to include a software and hardware operating time.
● The calculation of average values has been modified.
Manual structure
The manual has been divided into two volumes. The individual chapters ha ve been re vised
and some sections reformatted. All screens have been replaced.
1.5 Notational conventions
All statements for controlling monitoring operations (SM2 administration statements) and
the statement interfaces of the SM2U1 and SM2R1 utilities are fully supported by the SDF
statement syntax. ISP statements are no longer available.
The SM2 functions for selecting and controlling screen output can only be addressed via
ISP.
For a description of the SDF syntax, see the appendix.
U3585-J-Z125-8-76 7
README file Preface
1.6 README file
Information on any functional changes and additions to the current version of the product
described in this manual can be found in the product-specific README file. You will find the
README file on your BS2000 system under the file name SYSRME.SM2.130.E.
Ask your systems support staff for the user ID under which the README file is located. You
can view the README file by using the /SHOW-FILE command or an editor, and you can
print it out on a standard printer by means of the following command:
/PRINT-DOCUMENT SYSRME.SM2.130.E, LINE-SPACING=*BY-EBCDIC-CONTROL
8 U3585-J-Z125-8-76
2 The SM2 monitoring program
2.1 Overview
SM2 is a shareable program that runs under BS2000. It provides the user with inf ormation
on the status and performance of the system, making it possible to improve the performance of both the system and its applications. The mode and scope of monitored data
acquisition in SM2 can be controlled using commands and statements.
SM2 generates a number of system tasks which are responsible for collecting data and
writing it to the SM2 output files.
SM2 consists of privileged and nonprivileged components. Figure 1 shows the exchange of
information between the various SM2 tasks.
U3585-J-Z125-8-76 9
Overview SM2 monitoring program
SM2
administrator
BS2000
system kernel
SM2
global write
task
SM2
user
SM2
monitoring
task
SM2
buffer
...
SM2
user write
task
SM2
user
Monitored
task
Key:
Flow of
monitored data
Communication for
controlling monitoring
operations
Boundary between
privileged and
nonprivileged
program sections
SM2
file
Figure 1: SM2 configuration
User
output file
Privileged
SM2 code
Nonprivileged
SM2 code
10 U3585-J-Z125-8-76
SM2 monitoring program Operating modes/users
2.2 SM2 operating modes
SM2 offers three operating modes:
1. Real-time monitoring (information output to a data display terminal)
2. Background monitoring (information output to a file for subsequent analysis)
3. User-specific task monitoring
The operating modes can be used simultaneously.
When mode 1 is used, SM2 periodically outputs monitored data in the form of reports on
the data terminal. The color version of the 9752 Data Display Terminal is supported. This
data refers to the latest monitoring cycle (e.g. 150 seconds) and can be used as snapshots
for assessing the current system status. Only selected SM2 monitored data is output in this
mode.
When mode 2 is used, the collected data is output in the form records to a file, the SM2
output file. This data can be analyzed later using SM2 products. This mode allows for
selective use of the SM2 operands and is equally suitable for long-term monitoring with a
low system load and for brief special-purpose monitoring with a correspondingly higher
system load.
When mode 3 is used, the nonprivileged user (using the START-TASK-MEASUREMENT
command) can register tasks under his/her own ID for monitoring by SM2. In addition to the
task-specific monitored data, a command counter and SVC task statistics can be
requested. In the case of user-specific task monitoring, all information is written to userspecific SM2 output files.
U3585-J-Z125-8-76 11
Operating modes/users SM2 monitoring program
2.3 Users
Privileged users
Privileged SM2 users are users who have been assigned the system privilege
SW-MONITOR-ADMINISTRATION. These fall into the following categories: "primary"
administrators, "secondary" administrators and other privileged users.
Entry of an administration statement gives the first privileged SM2 user the status of SM2
primary administrator. This user retains this status until he/she terminates his/her program
or switches over to online analysis using the CALL-EVALUATION-PART statement. The
SM2 primary administrator has all privileges, and there can only be one primary administrator at any one time.
The primary administrator is the only person who is authorized to admit other (secondary)
administrators (MODIFY-ADMINISTRA T OR-ADMISSION statement) and to permit nonprivileged users to run monitoring programs (MODIFY-USER-ADMISSION).
With the exception of the MODIFY-ADMISSION statements, the secondary administrators
have the same rights as the primary administrator.
All administrators can:
– create and close the SM2 output file
– set monitoring parameters
– activate or deactivate optional monitoring runs.
In addition to these privileges, the SM2 administrators can also use all statements of a
nonprivileged user. All other privileged users are only offered the SHO W functions and the
SELECT-HOSTS statement in the administration section.
Notes
SM2 cannot prevent the various administrators from carrying out operations which
contflict with one another. It is only possible to make sensible use of the system if the
different administrators agree on a common course of action.
Modification of the variables to be monitored and functions which can be added may
influence the real-time monitoring of other SM2 users. Modification of the monitoring
cycle applies to all SM2 users.
Nonprivileged users
All users who call the SM2 routine from a user ID without the system privilege
SW-MONITOR-ADMINISTRATION are categorized as nonprivileged users. These users
are only authorized to use the statements for controlling output during real-time monitoring.
12 U3585-J-Z125-8-76
SM2 monitoring program Authorization table
2.4 Table of authorizations
The following table indicates which users are permitted to carry out which functions and
under what circumstances.
Sys. priv. SWMONADM
Function (group)
Start/stop functions
Start SM2 monitor
Terminate SM2 monitor
Administration functions
MODIFY-ADMINISTRATOR-ADMISSION
MODIFY-USER-ADMISSION
MODIFY-MEASUREMENT-PERIODS
OPEN-/CLOSE-LOG-FILE
ADD/REMOVE measurement object
SET/MOD measurement PARAMETER def.
INITIATE-COSMOS
START-/CHANGE-/STOP-MEASUREMENT-
PROGRAM
SHOW statements
CALL-EVALUATION-PART
SELECT-HOSTS
Analysis functions
OUTPUT / START
DEVICE
SELECT-DAB / -PCA / -PERIODIC-TASK / -PFA/
-UTM / -SYMMETRIX
STATUS
REPORT
RESTART
FILE
START-/CHANGE-/STOP-ISAM-STATISTICS
SHOW-USER-MEASURED-OBJECTS
CALL-ADMINISTRATION-PART
BS2000 commands
START-/STOP-TASK-MEASUREMENT 6 6 6 2 n
Other SM2 features
Monitor foreign files
Monitor foreign ISAM pools
Monitor foreign tasks
Monitor SVCs/PCounter
Table 1: Authorization table
Pr. adm Sec. adm
y
5
y
y
y
y
y
y
y
y
y
y
y
y
y
y
y
y
y
6
6
4
y
3
3
6
6
Other Priv. usr User
n
5
n
n
y
y
y
7
n
7
y
y
y
y
y
y
y
y
y
6
6
4
y
3
3
6
6
n
5
n
n
n
n
n
n
n
n
y
y
y
y
y
y
y
y
y
6
6
4
y
n
n
6
6
n
5
n
n
n
n
n
n
n
n
n
n
n
y
y
y
y
1
1
2
2
4
n
n
n
n
6
n
5
n
n
n
n
n
n
n
n
n
n
n
y
y
y
y
1
1
n
n
4
n
n
n
n
6
U3585-J-Z125-8-76 13
Activating the monitoring task SM2 monitoring program
Key
Pr. adm: Primary administrator
Sec. adm: Secondary administrator
Other: Other privileged user
Priv. user: Users permitted to execute a monitoring program as specified by MODIFY-
USER-ADMISSION
User: Users not permitted to execute a monitoring program
y: Yes
n: No
1: Yes, but some reports are reserved for privileged users
2: Yes if the corresponding monitoring process is permitted for the caller
3: Yes, but only for the corresponding privileged monitoring program
4: Yes, but only for objects registered by this user
5: System privilege SUBSYSTEM-MANAGEMENT required; SM2 privileges
are irrelevant
6: Yes, provided that "Priv. user" is also active
7: Yes, except for the COSMOS monitoring program
2.5 Activating a monitoring task
When SM2 is called for the first time during a session, the SM2 monitoring task is activated.
This is an internal task which collects the monitored data, edits it, and stores it in a central
buffer. F rom this buffer , the data is transf erred to the v arious SM2 user tasks or to the SM2
output file for output. If different offline and online periods are selected, the monitored data
is written to two central buffer stores.
The precise time at which the SM2 monitoring task was initiated and the online or offline
period selected is given in the SM2 MEASUREMENT STATUS under SM2 GATHERING
TASK CREATED AT, ONLINE PERIOD, OFFLINE PERIOD.
14 U3585-J-Z125-8-76
SM2 monitoring program Acquisition of monitored data
2.6 Acquisition of monitored data
SM2 records a wide range of monitored data and outputs it to either a screen and/or a file.
The data is collected at regular intervals, called monitoring cycles. Some of the monitoring
tasks are carried out by default, while others are carried out by special monitoring programs
which can be activated as and when required.
The monitored data is then processed and displayed internally by SM2 or by independent
SM2 monitoring and analysis routines.
There are three methods of collecting data:
● Based on the monitoring cycle
Most of the monitored data (e.g. the CPU load) is collected at the end of each
monitoring cycle. The current count (or time stamp) is taken from the SM2 or system
tables, and the difference between this value and the value at the end of the last
monitoring cycle is calculated.
The same procedure is used to calculate the intermediate values for the methods
described below at the end of the monitoring cycle (except for the monitoring programs
DISK and TASK).
A monitoring cycle can be set to between 10 seconds and 1 hour.
● Based on the sampling cycle
Because the status is constantly changing, it may not be appropriate to wait until the
end of the monitoring cycle to query some monitored data (e.g. device utilization). When
acquiring this type of monitored data, it is recommended that the monitoring cycle be
broken down into several sampling cycles. At the end of each sampling cycle, the
current monitored variable is obtained. Then, at the end of each monitoring cycle, the
mean value across all sampling cycles is computed.
A sampling cycle can be set between 200 milliseconds and 10 seconds.
● Event-driven
Some monitored data (e.g. the distribution of disk operation times) is determined by
monitoring events within the system. When an event occurs, such as the start of an
input/output operation, an SM2 routine is activated which collects the event-specific
data. This data is then used to calculate the monitored variables.
U3585-J-Z125-8-76 15
Acquisition of monitored data SM2 monitoring program
Analysis routine
SM2R1
SM2R1-PC
Output
buffer
S M 2
SM2
file
in BS2000
SM2 data
buffer
Monitoring programs Tables Counter BS2000 monitored object, event
Figure 2: Collecting monitored data
buffer
I/O
16 U3585-J-Z125-8-76
SM2 monitoring program Monitoring cycle
2.7 Monitoring cycle
As mentioned above, SM2 collects data at regular intervals known as monitoring cycles
(except for the monitoring programs DISK and TASK). At the end of each monitoring cycle,
the processed data is written to the data buffer and to the SM2 output file, if one is av ailable .
The length of the monitoring cycle is preset at 150 seconds when the subsystem is started.
The SM2 administrator can change this setting using the OFFLINE-PERIOD operand of the
MODIFY-MEASUREMENT-PERIODS statement. The value is entered under OFFLINE
PERIOD in SM2 MEASUREMENT STATUS .
Online monitoring cycle
The online cycle also defines a monitoring cycle in which the monitored data for online
analysis is collected and written to another buffer . The online cycle is deactivated by def ault.
It can be modified using the ONLINE-PERIOD parameter in the MODIFY-MEASUREMENT PERIODS statement. When the online cycle is deactivated, the ONLINE-PERIOD corresponds to the OFFLINE-PERIOD. The CYCLE column in each report shows the online
value currently set. In addition, the value under the ONLINE-PERIOD column is entered in
SM2 MEASUREMENT STATUS.
The online cycle should only be activated if the delay at the terminal is too great when
carrying out online monitoring. This may happen if a long monitoring cycle is set for output
to an SM2 output file. Please note that by activating the online cycle , you also increase the
workload on the system.
Sampling cycle
For monitored-data acquisition based on random sampling, SM2 subdivides the monitoring
cycle into a large number of small, regular time slots called sampling cycles. The monitoring
task is activated at the specified interval for data collection. At the end of each sampling
cycle, snapshots are taken for some v ariables. F rom the large number of snapshots, a mean
value is computed for each monitoring cycle.
The accuracy of the computed value thus depends on the length of the sampling cycle:
The shorter the sampling cycle, the greater the number of snapshots that can be taken
during the monitoring cycle. The SM2 workload on the system, however, increases simultaneously.
The sampling cycle is preset to 800 milliseconds when the subsystem is started.
The SM2 administrator can change this value in the SAMPLING-PERIOD operand of the
MODIFY-MEASUREMENT-PERIODS statement. The SAMPLING PERIOD column in the
SM2 MEASUREMENT STATUS indicates the current setting, while the SAMPLES column
in the individual reports contains the number of samples taken during the latest monitoring
cycle.
U3585-J-Z125-8-76 17
Outputting and saving monitored data SM2 monitoring program
To obtain data, SM2 counts and
val1
Monitoring cycle
Sampling cycle
val2 val3 val4 ..... valn
system counts are evaluated.
Mean values are computed,
percentage frequencies are
determined, and total values are
calculated in order to arrive at
monitored data.
n
1
value
=
Figure 3: Relationship between the sampling cycle and the monitoring cycle
val
-- -
∑
n
j
1
2.8 Outputting and saving monitored data
At the end of a monitoring cycle, the monitored data is summarized into groups which have
similar content, extended by the SM2 monitoring task to include additional inf ormation (time
stamp, host, ...), and then written to a central SM2 data buffer.
If a user has activated online monitoring, the monitored data is transferred from the central
data buffer to the b uffer of the user task. Note that sensitiv e data is transf erred to privileged
users only. The monitored data is then processed and output in the address space of the
user task.
For background monitoring, the SM2 administrator can open an SM2 output file. At the end
of a monitoring cycle, the monitored data is grouped into data records and written to the
SM2 output file. The information in the SM2 output file can be used for subsequent problem
and trend analyses. Special analysis routines are provided for this purpose.
18 U3585-J-Z125-8-76
SM2 monitoring program Outputting and saving monitored data
Methods used for compressed output of monitored data
1. Mean values
SM2 adds up the sampled values at the end of each sampling cycle and computes a
mean value for each monitoring cycle (e.g. the number of tasks) as follows:
Sum of sampled values
Average =
Number of samples
2. Percentage frequency
At the end of each sampling cycle, SM2 records the occurrence of specific events or
conditions and computes the percentage frequency of events (conditions) in relation to
the total number of samples (e.g. channel utilization) as follows:
Sum of occurrences
Frequency = *100 %
Number of samples
3. Ascending system counters and SM2 counters
The activity during the monitoring cycle can be determined by calculating the
differences between values gathered at the beginning and the end of the monitoring
cycle (e.g. idle time).
4. Computation
SM2 computes some data from monitored data, e.g. the summation values for all
categories.
Uninterruptible clock resetting
SM2 works internally with UTC time. During the switch from daylight saving time to winter
time and vice versa, SM2 can continue to work without interruption.
U3585-J-Z125-8-76 19
SM2 monitored variables SM2 monitoring program
2.9 SM2 monitored variables
The table below shows the main monitored variables of SM2.
A table showing the links between the monitored variables and monitoring programs,
ONLINE reports, and SM2R1 report groups can be found in chapter “Variables reports” on
page 361.
Monitored variable Description
Cache
Access statistics and hit rates for
DAB caches
Access statistics for ISAM pools Number of ISAM accesses per second to pages in the ISAM
Access statistics and hit rates for
hiperfiles (PFA)
Access statistics and hit rates for
SYMMETRIX controllers
Access statistics and hit rates for
controllers with cache (PCA)
CPU
CPU utilization Time during which the processor is in one of the following
Number of system calls Number of SVC calls in TU/TPR per second, and the sum of all
Files
Catalog access statistics Number of read/write accesses to catalog entries/JV entries of
Table 2: SM2 monitored variables
Number of read/write accesses to DAB cache areas per second
and the percentage frequency of read/write accesses to DAB
subareas without disk access.
pool and directly to disk, as well as the number of ISAM
accesses with a delay per second.
Size of the ISAM pool, reserved pages and fixed pages in PAM
pages.
Number of read/write accesses to PCA or DAB cache areas per
second, and their percentage share of the total number of read/
write accesses.
Number of failed attempts to use the cache per second.
Number of read/write accesses to SYMMETRIX controllers and
devices.
Proportion of read/write accesses handled successfully by the
cache.
Number of read/write accesses and their percentage share of
the total number of read/write accesses for PCA controllers,
cache areas, and devices.
statuses: TU, TPR, SIH, IDLE or STOP.
In SR2000 systems, /390 time will also be output.
SVC calls per second.
files per second, as well as average access times in milliseconds.
20 U3585-J-Z125-8-76
SM2 monitoring program SM2 monitored variables
Monitored variable Description
Files (continued)
File access statistics Number of PAM-WAIT, PAM-CHECK-READ, PAM-WRITE
operations and input/output operations per second for one file.
Distribution of input/output operations to files for selected disks.
File access times Average access time in milliseconds for each input/output.
IOs
Number of input/output operations Number of input/output operations per second.
Number of paging input/output
operations
Channel utilization and channel
transmission rates
Device utilization and transmission
rates
Length of device queues Number of input/output requests addressed to a device.
Duration of input/output operations Hardware operating time between SDV command and channel
Access distribution to cylinders/
PAM blocks of disks
Number of input/output operations Number of input/output operations per second.
Communication
Response times Aver age time in seconds between the receipt of a message on
Transaction times Average time in seconds between input and the last output.
Thinking time Time between an output and the resulting subsequent input.
Wait time Average wait time for incoming messages before being
Transaction rate Number of transactions per second.
Length of transaction message Average length in bytes of input/output messages for selected
Number of input/output operations
of network devices
Network transmission rate Transmission rate for all TCP/IP connections; data transfer
Table 2: SM2 monitored variables
Number of input/output operations or transferred PAM blocks
per channel.
Device utilization in % without/because of paging activities.
interruption for input/output operations of a specific device,
measured in milliseconds.
Cylinder number/PAM block addressed for input/output
operation.
the system and the sending of a response to the application.
processed by the application.
connection sets.
Number of read/write accesses per second, and the number of
bytes transferred.
information for specific connections.
U3585-J-Z125-8-76 21
SM2 monitored variables SM2 monitoring program
Monitored variable Description
Memory
Main memory utilization Number of 4-KB pages in main memory.
Utilization of paging area Number of pageable 4-KB pages on the devices.
Utilization of virtual address space Number of class 1 through class 4 pages in the virtual address
space.
Page fault rate Number of page fault interrupts per second.
Access statistics forglobal storage Number of read/write accesses to global storage per second
and the number of bytes transferred per second.
Subsystem
PCS data Service rate of the affected categories in %, job delays and
SERVICE RATEs per second.
UTM response times and
transaction rates
Number and duration of send jobs
via MSCF
POSIX data File accesses, message accesses, semaphore activities, buff er
Lock requests to the DLM Number of enqueue, convert, dequeue and information lock
Data on synchronization functions
in HIPLEX (NSM)
HSMS data Data on the migration of files to the background level and the
Task
Number of tasks Number of BATCH, DIALOG, TP and system tasks.
Length of task queues Number of tasks in a category.
Dwell times in task queues Dwell times in task queues per category.
Frequency of task activation and
deactivation
Task-specific utilization of
resources
Utilization and queues for task
locks
Number of dialog and asynchronous transactions per second
and the average time in seconds taken for the transaction.
Number of send jobs per second, as well as average total and
wait times.
utilization and all types of system accesses per second.
requests per second, as well as the number of lock allocations
and releases per second.
retrieval of files to the processing level.
Service units per second, percentage CPU utilization,
input/output operations per second, used pages in 4-KB pages,
paging read per second.
Percentage utilization of a task lock and the n umber of tasks in
the task lock queue.
Table 2: SM2 monitored variables
22 U3585-J-Z125-8-76
SM2 monitoring program SM2 monitored variables
Monitored variable Description
VM2000
Hypervisor activities Percentage active and idle times of the system.
Guest system activities Planned and measured CPU utilization as a percentage.
Table 2: SM2 monitored variables
U3585-J-Z125-8-76 23
Analyzing monitored data SM2 monitoring program
2.10 Analyzing monitored data
There are several tools on various different platforms which can be used to process and
display monitored data.
SM2-ONLINE (BS2000/OSD)
SM2-ONLINE is a component of the SM2 monitoring program. At the end of each
monitoring cycle, it outputs monitored data in the form of reports. For further information,
please refer to chapter “SM2 screen output” on page 201ff.
SM2R1 (BS2000/OSD)
SM2R1 is a component of SM2 and analyzes the SM2 output file. SM2R1 outputs statistics
in tabular form and time series in chart form to a file which is ready to print. For further information, please refer to the chapter “SM2R1 analysis routine” in Volume 2 of the SM2
manual [21].
SM2R1-PC (Microsoft Windows)
The SM2R-PC routine can be purchased separately, and is a development of SM2R1. It
makes use of the graphical options off ered by a PC when displaying monitored data. SM2
output files from various BS2000/OSD computers can be analyzed in an single analysis
run. The result data can be displayed in tabular f orm and can be edited graphically. Charts
can be copied into the clipboard and transferred to any Windows application that can
access the clipboard. For further information, please refer to the chapter “Additional utility
routines” in Volume 2 of the SM2 manual [21].
SM2ONLINE-PC (Microsoft Windows)
The SM2ONLINE-PC utility can be purchased separately and is used to simultaneously
display and monitor SM2 data from sev eral BS2000/OSD computers online. The monitored
data is supplied in the form of charts which are updated at the end of every monitoring cycle.
Each monitored variable can be subjected to threshold monitoring. Alarms signalling
threshold violations can be both audible and visual. For further information, please refer to
the chapter “Additional utility routines” in Volume 2 of the SM2 manual [21].
SM2-PA (BS2000/OSD)
The SM2-P A routine can be purchased separately and is used to analyze user-specific SM2
output files from user task monitoring operations. The result data is displayed in the f orm of
statistics and offers information on task resource utilization and/or the performance of user
programs. For further information, please refer to the chapter “Additional utility routines” in
Volume 2 of the SM2 manual [21] and the manual “SM2-PA” [8].
24 U3585-J-Z125-8-76
SM2 monitoring program Performance expectations of the user
2.11 Performance expectations of the user
When assessing the performance of a DP system, the user is generally interested in:
● the time it takes for the system to respond to requests,
● how many requests the system can handle within a given period of time, and
● what resources are used.
The performance level of the hardware (e.g. the number of instructions processed per
second) and of the operating system (e.g. task management efficiency) are generally of
secondary interest.
Interactive mode criteria
A distinction is made between inquiry-and-transaction mode and timesharing mode:
In inquiry-and-transaction mode, the terminal users can communicate only with
programs which are specific to a given application. Normally, a large number of users
interact with a relatively small number of application programs.
In interactive (or timesharing) mode, the terminal users enter their own application
programs for interactiv ely processing a giv en task. The application prog rams are generally
controlled by system programs for creating, testing and updating files and programs.
In both inquiry-and-transaction mode and timesharing mode, the unit of DP operation is a
transaction.
The transaction time is the delay between the arrival of the input at the host and the output
of the acknowledgment. The DP system can output several responses with different
response times for a singletransaction. If resources are not released when a transaction is
completed but are reserved for the ne xttransaction, the transaction is known as a multi-step
transaction.
U3585-J-Z125-8-76 25
Performance expectations of the user SM2 monitoring program
The sum of single-step and multi-step transactions required for handling one application is
called an operation.
The characteristic criteria for describing performance expectations are:
● Transaction rate
Sum of all successfully completed transactions per unit of time
● Response time
Time required for processing by the DP system
● Number of simultaneously active users.
The monitored values for these characteristics are determined in SM2 by the monitoring
program BCAM-CONNECTION (page 35), RESPONSETIME (page 45) and UTM
(page 53).
Batch processing criteria
In batch processing, the unit of DP processing is a job.
● Throughput rate
Number of successfully processed jobs per unit of time
● Dwell time
Time required for processing a job
For the individual user, a satisfactory performance (short response time, short dwell time)
is obtained if the required resources are available when they are requested. This
requirement can easily be meet if utilization of the requested resources (in particular of the
input/output devices) is kept to a minimum.
The DP system performance, on the other hand, is characterized by the transaction and
throughput rates, where maximum utilization of the resources is desirable for economic
reasons.
These contradictory performance requirements relating to resource utilization can be
fulfilled only by carefully planning the workload and the associated resource requirements.
26 U3585-J-Z125-8-76
SM2 monitoring program System monitoring
2.12 Using SM2 for system monitoring
Depending on the mode of monitored-data acquisition (frequency, scope), we must
consider two different applications:
● Trend monitoring (= long-term monitoring) for obtaining data for system capacity
planning
● Bottleneck analysis for locating and eliminating performance problems
Trend monitoring
The utilization data of the following resources is required in order to carry out long-term
system capacity planning:
– CPUs
– channels
– devices
– main memory
Additional monitoring routines need not be activated.
It is advisable to use the following settings for monitoring periods:
Sampling cycle (SAMPLING-PERIOD): 1000 milliseconds
Monitoring cycle (OFFLINE-PERIOD): 5 minutes
Analysis subinterval 1 hour
The monitoring period should cover the entire period from SYSTEM READY through to
SHUTDOWN. If output of the online screen report takes too long during the session, you
can shorten the online monitoring cycle.
Monitoring times are set using the MODIFY-MEASUREMENT-PERIODS statement.
It is a good idea to create a new SM2 output file every day (OPEN-LOG-FILE / CLOSELOG-FILE statements). The SM2U1 routine can be used to combine (and split) daily SM2
output files to create one large file, known as the master SM2 output file. The daily SM2
output files must be added to the master SM2 output file in chronological order.
U3585-J-Z125-8-76 27
System monitoring SM2 monitoring program
Bottleneck analysis
Before monitoring is started, you must clarify any performance problems, i.e. performance
expectations that are not satisfied. The following problems may exist:
● System-oriented performance problems
These arise if the system throughput rate is unsatisfactory, and are indicated by a low
transaction rate and/or throughput rate. The most likely cause is the o verloading of one
or more resources.
● User-oriented performance problems
These occur due to long delays when handling specific load requirements.
The following monitored v ariables should be used to analyze bottlenec ks. SM2 also allo ws
for more extensive analysis through addition monitoring programs.
These monitored variables and monitoring programs make it easier to locate overloaded
resources:
Monitored variable Monitoring program
Number of tasks in the system queues and at devices Monitored by default
Number of input/output operations per device Monitored by default
Working set per category Monitored by default
CPU utilization and number of input/output operations per category SYSSTAT
Number of input/output operations and volume of data transf erred per
channel
Access to catalog entries CMS
Number of transactions RESPONSETIME, BCAM-
Table 3: Monitoring programs used to locate overloaded resources
CHANNEL-IO
CONNECTION, and UTM
The following settings are recommended for monitoring times (MODIFY-MEASUREMENTPERIODS statement):
Sampling cycle (SAMPLING PERIOD): 400 milliseconds
Monitoring cycle (OFFLINE-PERIOD): 60 seconds
Analysis subinterval: 1 - 5 minutes
Monitoring period: 0.5 - 5 hours
Monitoring must be carried out during peak load periods.
Due to the shorter monitoring cycle and the activated monitoring programs, bottleneck
analysis produces a large volume of data compared to trend monitoring. The volume of data
corresponds to the number of objects monitored and the number of events (DISK
monitoring program). The resulting SM2 output file may be very large.
28 U3585-J-Z125-8-76
SM2 monitoring program Efficient DP system usage
Because of the volume of data generated, it does not make sense to copy all data record
types into the master SM2 output file. SM2U1 can be used to suppress certain data records
when updating the master SM2 output file.
To investigate delays when handling special load requirements, you will need further information in addition to the system utilization data described above. To begin with, the
monitoring program PERIODIC-T ASK or TASK can be used to select a task. The DISK-FILE
monitoring program can be used for ov erloaded disks to determine the files accessed most
frequently. It is not possible to list general guidelines for the additional selection of
monitoring programs. For further information, please refer to the “Performance Handbook”
[5].
2.13 Basic procedure for ensuring efficient DP system usage
T o minimiz e performance analysis problems, the f ollowing procedure should be adhered to:
● Define the performance expectations (see page 25).
● Check the extent to which performance expectations are satisfied once productive
operation has begun. This involves carrying out a “basis” test.
● If some performance expectations are not satisfied, check whether these are system-
oriented or user-oriented.
● Check for errors in the system setup.
● Carry out a bottleneck analysis (see the section “Automatic performance analysis” in
the chapter “SM2R1 analysis routine” and the description of the START-AUTOMATICANALYSIS statement in Volume 2 of the SM2 manual [21]).
Concentrate on those bottlenecks whose elimination promises the greatest performance improvement.
● After eliminating the detected bottlenecks, repeat the bottleneck analysis. Man y bottle-
necks remain hidden until after some kind of intervention.
● Monitor the system at regular intervals (see page 27) to detect saturation symptoms in
the main resources (due to increasing loads) and to avoid critical system conditions.
The procedure is described in detail in the “Performance Handbook” [5], which also e xplains
how to interpret the results.
U3585-J-Z125-8-76 29
SM2 in computer networks SM2 monitoring program
2.14 SM2 in computer networks
SM2 provides functions for the central online monitoring and control of monitored-data
acquisition in a computer network.
The monitored data is independently recorded on each host in the network by a local SM2
monitoring program, and is exchanged between the hosts via a LAN. In this wa y, the online
reports of all hosts can be output on any computer in the network.
There are also reports which output the combined monitored data of the various hosts.
All control statements for the acquisition of monitored data (changing monitoring cycles,
switching monitoring programs, etc.) can be entered on any host for all computers in the
network.
The SM2 functions for monitored-data acquisition in a network are also available in a
HIPLEX network.
Prerequisites for using SM2 in computer networks
1. An MSCF connection of type CCS (Closely Coupled System) must exist between all
hosts in the network.
For further information on this connection type, refer to the “HIPLEX MSCF” manual [4].
2. The SM2 functions for monitored-data acquisition in a network can only be used for
those computers in the network which run the same version of SM2.
3. SM2 must be started once on all computers in the network (e.g. with ST AR T-SM2) or at
least loaded via the DSSM command /START-SUBSYSTEM
SUBSYSTEM-NAME=SM2.
Selecting hosts with SM2
The privileged statement SELECT-HOSTS (see page 110) selects the computers to which
all subsequently entered SM2 statements are to be sent. This statement always applies to
the user who issues it. Following the END statement and a restart, the default value applies
again, i.e. SM2 statements are only executed on the local host.
The SHOW-SELECTED-HOSTS statement (see page 140) provides information on the
selected hosts. This includes the time of the last monitored variable queried.
30 U3585-J-Z125-8-76
Loading...