![[Digital logo]](../../IMAGES/DIGITAL-LOGO.GIF){kind=logo}
![[HR]](../../IMAGES/REDBAR.GIF)
The command in this example requests that kilobyte map rate statistics collected for SCS be displayed for each node in the cluster. The display shows block transfer map activity between the node CURLEY and the hierarchical storage controller (HSC) SHEMP. Note that each node in the cluster is identified by its SCS node name.
The MONITOR STATES command initiates monitoring of the PROCESS STATES class, which shows the number of processes in each of the 14 scheduler states.
MONITOR STATES
/qualifier[,...]
One or more qualifiers as described in the Command Qualifier Descriptions section.
/ALL
Specifies that a table of all available statistics (current, average, minimum, and maximum) is to be included in the display and summary output. For summary output, this qualifier is the default for all classes; otherwise, it is the default for all classes except CLUSTER, MODES, PROCESSES, STATES, SYSTEM, and VECTOR./AVERAGE
Selects average statistics to be displayed in a bar graph for display and summary output./CURRENT
Selects current statistics to be displayed in a bar graph for display and summary output. The /CURRENT qualifier is the default for the CLUSTER, MODES, STATES, SYSTEM, and VECTOR classes./MAXIMUM
Selects maximum statistics to be displayed in a bar graph for display and summary output./MINIMUM
Selects minimum statistics to be displayed in a bar graph for display and summary output./PERCENT
/NOPERCENT (default)
Controls whether statistics are expressed as percent values in display and summary output. The /PERCENT qualifier is applicable only to the DISK, MODES, SCS, and STATES class names.
The STATES class shows the number of processes in each of the 14 scheduler states, as follows:
- Collided Page Wait (COLPG)---Waiting for a faulted page in transition.
- Mutex & Miscellaneous Resource Wait (MWAIT)---Waiting for the availability of a mutual exclusion semaphore or a dynamic resource. The following is a summary of Mutex and Miscellaneous Resource Wait states and identifying codes, as they appear in the PROCESSES class display:
MWAIT Reason for Wait MUTEX Mutual exclusion semaphore RWAST AST wait (wait for system or special kernel AST) RWBRK Breakthrough (wait for broadcast message) RWCAP CPU capability required RWCLU Cluster state transition wait RWCSV Cluster server RWIMG Image activation lock RWLCK Lock database RWMBX Mailbox full RWMPB Modified page writer busy RWMPE Modified page list empty RWNPG Nonpaged dynamic memory RWPAG Paged dynamic memory RWPGF Page file full RWQUO Job quota RWSCS System Communications Services wait RWSNP System snapshot RWSWP Swap file space - Common Event Flag Wait (CEF)---Waiting for some combination of event flags to be set in a common event block.
- Page Fault Wait (PFW)---Waiting for a page to be read as a result of a page fault; resident processes.
- Local Event Flag Wait (LEF)---Waiting for one or more local event flags to be posted; resident processes.
- Local Event Flag (Outswapped) (LEFO)---Waiting for one or more local event flags to be posted; outswapped processes.
- Hibernate (HIB)---Hibernating, or process has executed a hibernate request; resident processes.
- Hibernate (Outswapped) (HIBO)---Hibernating, or process has executed a hibernate request; outswapped processes.
- Suspended (SUSP)---Process has executed a suspend request; resident processes.
- Suspended (Outswapped) (SUSPO)---Process has executed a suspend request; outswapped processes.
- Free Page Wait (FPW)---Waiting for a free page of memory.
- Compute (COM)---Ready to use the processor; resident processes.
- Compute (Outswapped) (COMO)---Ready to use the processor; outswapped processes.
- Current Process (CUR)---Using the processor.
The data items can also be displayed as percentages of all processes.
Note that the Current Process is always the process running MONITOR, because MONITOR is running when each measurement is made.
For performance reasons, MONITOR does not synchronize the scanning of process state data structures with operating system use of those structures. It is therefore possible that MONITOR will display certain anomalous state indications.
$ MONITOR/INPUT/SUMMARY/NODISPLAY -
_$/BEGINNING=29-APR-1996:13:00 -
_$/ENDING=29-APR-1996:14:00 STATES/PERCENT/ALL
$ TYPE MONITOR.SUM
OpenVMS Monitor Utility
PROCESS STATES (%)
on node SAMPLE From: 29-APR-1996 13:00:00
SUMMARY To: 29-APR-1996 14:00:00
CUR% AVE% MIN% MAX%
Collided Page Wait 0.0 0.0 0.0 0.0
Mutex & Misc Resource Wait 0.0 0.0 0.0 0.0
Common Event Flag Wait 0.0 0.0 0.0 0.0
Page Fault Wait 4.3 1.4 0.0 4.3
Local Event Flag Wait 34.7 31.7 34.7 42.8
Local Evt Flg (Outswapped) 0.0 9.0 0.0 19.4
Hibernate 43.4 40.7 43.4 52.1
Hibernate (Outswapped) 0.0 4.3 0.0 15.4
Suspended 0.0 0.0 0.0 0.0
Suspended (Outswapped) 0.0 0.0 0.0 0.0
Free Page Wait 0.0 0.0 0.0 0.0
Compute 13.0 7.3 4.3 13.0
Compute (Outswapped) 0.0 0.8 0.0 3.2
Current Process 1.0 1.0 1.0 1.0
PLAYBACK SUMMARIZING
The commands in this example generate and display a PROCESS STATES summary. Note that since use of the Return key is not permitted within a single MONITOR command following the MONITOR> prompt, the MONITOR command is entered at DCL level. The summary shows that, on the average, 14.1 percent of processes were swapped out for the summarized period. Note that the summary was requested for data covering only the hour between 1 P.M. and 2 P.M., although the input file could have contained data covering a longer period.
The MONITOR SYSTEM command initiates monitoring of the SYSTEM statistics class, which shows several of the most important items from other classes.
MONITOR SYSTEM
/qualifier[,...]
One or more qualifiers as described in the Command Qualifier Descriptions section.
/ALL
Specifies that a table of all available statistics (current, average, minimum, and maximum) is to be included in the display and summary output. For summary output, this qualifier is the default for all classes; otherwise, it is the default for all classes except CLUSTER, MODES, PROCESSES, STATES, SYSTEM, and VECTOR./AVERAGE
Selects average statistics to be displayed in a bar graph for display and summary output./CURRENT
Selects current statistics to be displayed in a bar graph for display and summary output. The /CURRENT qualifier is the default for the CLUSTER, MODES, STATES, SYSTEM, and VECTOR classes./MAXIMUM
Selects maximum statistics to be displayed in a bar graph for display and summary output./MINIMUM
Selects minimum statistics to be displayed in a bar graph for display and summary output.
Because the SYSTEM class collects the most significant performance statistics from other classes in a single display, it is particularly useful to system managers and other users seeking a general overview of system activity. The SYSTEM class includes the following data items:
- Interrupt Stack (on VAX systems) or Interrupt State (on Alpha systems)
- MP Synchronization
- Kernel Mode
- Executive Mode
- Supervisor Mode
- User Mode
- Compatibility Mode (meaningful on VAX systems only)
- Idle Time
- Process Count
- Page Fault Rate
- Page Read I/O Rate
- Free List Size
- Modified List Size
- Direct I/O Rate
- Buffered I/O Rate
The following two display formats are provided, depending on the classname qualifier specified:
- A tabular style format for the /ALL qualifier
- A bar graph style format for the /AVERAGE, /CURRENT, /MAXIMUM, and /MINIMUM qualifiers
Note that the bar graph version of the SYSTEM class (shown in Example 2) contains the following data, which differs from the tabular version:
- All of the CPU processor modes except Idle Time are summarized in the CPU Busy segment.
- In the Page Fault segment, the page read I/O rate is indicated by a vertical bar. The bar provides a visual estimate of the proportion of the total page fault rate that caused read I/O operations (the hard fault rate). The hard fault rate appears to the left of the bar.
- Four segments show the processes that are currently the top consumers of CPU (since the last screen update), page faults, direct I/Os, and buffered I/Os.
Any process that MONITOR designates as a top user process must be swapped in at the beginning and ending of the display interval or at the beginning and ending of the entire period covered by a summary.
When the lower bar graph (top user) and the corresponding upper bar graph (overall system measure) are tracking the same statistic for the same interval (as in Example 2), it is reasonable to compare the two graphs. This will be the case in the following situation:
- SYSTEM is the only class being monitored (no other class names have been specified with the MONITOR command).
- The CURRENT statistic is specified.
- The /INTERVAL and /VIEWING_TIME values are equal.
Otherwise, you should exercise care in making such comparisons because the top user statistic is always CURRENT, while the overall system measure statistic may be CURRENT, AVERAGE, MAXIMUM, or MINIMUM.
Rates for top users are calculated based on the interval between two successive screen displays, while overall system rates are based on the collection interval. These two interval values can be different whenever one or more classes are being monitored with the SYSTEM class, or when /INTERVAL and /VIEWING_TIME values differ.
While other upper boundary figures for the SYSTEM class bar graphs are constants, the figures for Free List Size and Modified List Size are derived from the physical memory configuration and system parameters of individual systems. The upper boundary figure for the Free List is the number of pages available after deducting the pages permanently allocated to the operating system. This figure, sometimes referred to as balance set memory, is the number of pages that can be allocated to processes, the Free List, and the Modified List. The upper boundary figure for the Modified List is the value of the MPW_HILIMIT system parameter. Note that both upper boundary figures are calculated when the MONITOR request is initiated and do not change thereafter.
Examples
MONITOR> MONITOR SYSTEM/ALL
OpenVMS Monitor Utility
SYSTEM STATISTICS
on node SAMPLE
29-APR-1996 12:43:28
CUR AVE MIN MAX
Interrupt Stack 0.33 0.33 0.33 0.33
MP Synchronization 0.00 0.00 0.00 0.00
Kernel Mode 0.16 0.16 0.16 0.16
Executive Mode 0.00 0.00 0.00 0.00
Supervisor Mode 0.00 0.00 0.00 0.00
User Mode 0.50 0.49 0.50 0.50
Compatibility Mode 0.00 0.00 0.00 0.00
Idle Time 99.00 98.67 99.00 99.00
Process Count 14.00 14.00 14.00 14.00
Page Fault Rate 0.33 0.33 0.33 0.33
Page Read I/O Rate 0.00 0.00 0.00 0.00
Free List Size 4255.00 4255.00 4255.00 4255.00
Modified List Size 105.00 105.00 105.00 105.00
Direct I/O Rate 0.00 0.00 0.00 0.00
Buffered I/O Rate 0.16 0.16 0.16 0.16
MONITOR> MONITOR SYSTEM
Node: SAMPLE OpenVMS Monitor Utility 29-APR-1996 12:38:48
Statistic: CURRENT SYSTEM STATISTICS
Process States
+ CPU Busy (100) -+ LEF: 7 LEFO: 0
|**************************| HIB: 11 HIBO: 0
CPU 0 +--------------------------+ 100 COM: 4 COMO: 0
|******** | PFW: 0 Other: 1
+--------------------------+ MWAIT: 0
Cur Top: ASSEM_LIBRTL (34) Total: 23
+ Page Fault Rate (25) -+ + Free List Size (4604) -+
|**|*** | |********* | 12K
MEMORY 0 +--------------------------+ 100 0 +--------------------------+
|** | |******* | 500
+--------------------------+ + Modified List Size (146) +
Cur Top: ASSEM_SYS (11)
+ Direct I/O Rate (15) -+ + Buffered I/O Rate (0) -+
|****** | | |
I/O 0 +--------------------------+ 60 0 +--------------------------+ 150
|**** | | |
+--------------------------+ +--------------------------+
Cur Top: ASSEM_SYS (10) Cur Top: Cerb Servant (0)
The MONITOR TRANSACTION command initiates monitoring of the TRANSACTION class, which shows information about transactions coordinated by DECdtm services.
MONITOR TRANSACTION
/qualifier[,...]
One or more qualifiers as described in the Command Qualifier Descriptions section.
/ALL
Specifies that a table of all available statistics (current, average, minimum, and maximum) is to be included in the display and summary output. For summary output, this qualifier is the default for all classes; otherwise, it is the default for all classes except CLUSTER, MODES, PROCESSES, STATES, SYSTEM, and VECTOR./AVERAGE
Selects average statistics to be displayed in a bar graph for display and summary output./CURRENT
Selects current statistics to be displayed in a bar graph for display and summary output. The /CURRENT qualifier is the default for the CLUSTER, MODES, STATES, SYSTEM, and VECTOR classes./MAXIMUM
Selects maximum statistics to be displayed in a bar graph for display and summary output./MINIMUM
Selects minimum statistics to be displayed in a bar graph for display and summary output.
The TRANSACTION class consists of the following data items:
- Start Rate---The rate at which new transactions are started on the local node.
- Prepare Rate---The rate at which transactions on the local node are placed in the Prepared state by DECdtm services.
- One-Phase Commit Rate---The rate at which transactions on the local node complete using the one-phase commit operation. This operation, which consumes significantly fewer system resources, is used when there is only a single resource manager participating in the transaction.
- Total Commit Rate---The rate at which transactions on the local node are committed. This value is the combined total of one-phase and two-phase commit transactions.
- Abort Rate---The rate at which transactions on the local node are aborted.
- End Rate---The rate at which transactions that were started on the local node are committed.
- Remote Start Rate---The rate at which transaction branches are started on the local node.
- Remote Add Rate---The rate at which transaction branches are added on the local node.
- Completion Rate---The rate at which transactions complete, indexed by their duration in seconds. The following is a list of the Completion Rate categories:
Completion Rate 0--1 The number of transactions completed in 0--1 seconds (1 second or less) Completion Rate 1--2 The number of transactions completed in 1--2 seconds Completion Rate 2--3 The number of transactions completed in 2--3 seconds Completion Rate 3--4 The number of transactions completed in 3--4 seconds Completion Rate 4--5 The number of transactions completed in 4--5 seconds Completion Rate 5+ The number of transactions that took more than 5 seconds to complete
For example, a transaction that completes in 0.5 second is included in the statistics displayed for the Completion Rate 0--1 category.
Examples
MONITOR> MONITOR TRANSACTION/ALL
OpenVMS Monitor Utility
DISTRIBUTED TRANSACTION STATISTICS
on node SAMPLE
16-JAN-1996 14:52:34
CUR AVE MIN MAX
Start Rate 34.76 34.76 34.76 34.76
Prepare Rate 33.77 33.77 33.77 33.77
One Phase Commit Rate 0.00 0.00 0.00 0.00
Total Commit Rate 35.09 35.09 35.09 35.09
Abort Rate 0.00 0.00 0.00 0.00
End Rate 35.09 35.09 35.09 35.09
Remote Start Rate 31.12 31.12 31.12 31.12
Remote Add Rate 31.45 31.45 31.45 31.45
Completion Rate 0-1 35.09 35.09 35.09 35.09
by Duration 1-2 0.00 0.00 0.00 0.00
in Seconds 2-3 0.00 0.00 0.00 0.00
3-4 0.00 0.00 0.00 0.00
4-5 0.00 0.00 0.00 0.00
5+ 0.00 0.00 0.00 0.00
MONITOR> MONITOR TRANSACTION/MAXIMUM
OpenVMS Monitor Utility
+-----+ DISTRIBUTED TRANSACTION STATISTICS
| MAX | on node SAMPLE
+-----+ 16-JAN-1996 14:51:04
0 25 50 75 100
+ - - - - + - - - - + - - - - + - - - - -+
Start Rate 35 |**************
Prepare Rate 37 |**************
One Phase Commit Rate |
Total Commit Rate 35 |**************
Abort Rate |
End Rate 35 |**************
Remote Start Rate 33 |*************
Remote Add Rate 32 |************
| | | | |
Completion Rate 0-1 35 |**************
by Duration 1-2 |
in Seconds 2-3 |
3-4 |
4-5 |
5+ |
+ - - - - + - - - - + - - - - + - - - - -+
On VAX systems, the MONITOR VBS command displays information about the processing of virtual balance slots (VBS).
MONITOR VBS
The MONITOR VBS command includes the following data items:
- RBS Fault Rate---Rate at which processes fault from virtual balance slots into real balance slots (RBS)
- CPU Utilization---Amount of CPU time used by the operating system to support the virtual balance slot feature (at a rate of 10-millisecond clock ticks per second)
To see which processes have the highest RBS fault rates, issue the MONITOR PROCESSES/TOPRBS command.
MONITOR> MONITOR VBS
OpenVMS Monitor Utility
VIRTUAL BALANCE SLOT STATISTICS
on node SAMPLE
29-APR-1996 12:43:28
CUR AVE MIN MAX
RBS Fault Rate 4.62 3.80 0.33 7.61
CPU Utilization 0.99 0.24 0.00 0.99
This example shows the tabular style format for the VBS display.
The MONITOR VECTOR command displays the number of 10-millisecond clock ticks per second in which one or more vector consumers have been scheduled on each currently configured vector processor in the system.
MONITOR VECTOR
/qualifier[,...]
One or more qualifiers as described in the Command Qualifier Descriptions section.
/ALL
Specifies that a table of all available statistics (current, average, minimum, and maximum) is to be included in the display and summary output. For summary output, this qualifier is the default for all classes; otherwise, it is the default for all classes except CLUSTER, MODES, PROCESSES, STATES, SYSTEM, and VECTOR./AVERAGE
Selects average statistics to be displayed in a bar graph for display and summary output./CURRENT
Selects current statistics to be displayed in a bar graph for display and summary output. The /CURRENT qualifier is the default for the CLUSTER, MODES, STATES, SYSTEM, and VECTOR classes./MAXIMUM
Selects maximum statistics to be displayed in a bar graph for display and summary output./MINIMUM
Selects minimum statistics to be displayed in a bar graph for display and summary output.
The MONITOR VECTOR command displays the number of 10-millisecond clock ticks per second in which one or more vector consumers have been scheduled on each currently configured vector processor in the system. Because the operating system schedules vector consumers only on those processors identified as "vector present," the VECTOR class output never displays vector CPU time for those processors that are "vector absent."Note that, because vector consumers can use either the vector CPU, the scalar CPU, or both components of a vector-present processor, the vector CPU time in the VECTOR class display is not a strict measure of the actual usage of the processor's vector CPU component. Rather, it indicates the time during which a scheduled vector consumer has reserved both vector CPU and scalar CPU components of the vector-present processor for its own exclusive use.
The VECTOR class consists of the data item Vector Scheduled Rate, which is represented by a display of statistics that show the rates of 10-millisecond clock ticks per second during which vector consumers have been scheduled on each vector-present CPU.
MONITOR> MONITOR VECTOR
OpenVMS Monitor Utility
VECTOR PROCESSOR STATISTICS
+-----+ on node SAMPLE
| CUR | 12-JUN-1996 22:52:42
+-----+
Vector Consumers Scheduled 0 25 50 75 100
+ - - - - + - - - - + - - - - + - - - - -+
Vector Present CPU ID 0 13|*****
Vector Absent CPU ID 1 |
Vector Absent CPU ID 2 |
Vector Present CPU ID 4 58|**********************
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
+ - - - - + - - - - + - - - - + - - - - -+
This example shows the VECTOR class display for a multiprocessing system containing two vector-present processors, CPU 0 and CPU 4. Displayed statistics represent rates of 10-millisecond clock ticks per second. For an average of 13 ticks per second over the last collection interval, vector consumers have been scheduled on CPU 0. For an average of 58 ticks per second over the last collection interval, vector consumers have been scheduled on CPU 4.
6048P037.HTM OSSG Documentation 26-NOV-1996 12:43:14.98
Copyright © Digital Equipment Corporation 1996. All Rights Reserved.