| Previous | Contents | Index |
When you change the ACP/XQP cache parameters, remember to reboot the
system to make the changes effective. For more information about these
parameters, see Appendix K.
MONITOR IO
The MONITOR IO command initiates monitoring of the I/O class.
MONITOR IO
/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
Specifies that a bar graph of average statistics is to be included in the display and summary outputs./CURRENT
Specifies that a bar graph of current statistics is to be included in the display and summary outputs. The /CURRENT qualifier is the default for the CLUSTER, MODES, STATES, SYSTEM, and VECTOR classes./MAXIMUM
Specifies that a bar graph of maximum statistics is to be included in the display and summary outputs./MINIMUM
Specifies that a bar graph of minimum statistics is to be included in the display and summary outputs.
The IO class includes the following data items:
Data Item Description Direct I/O Rate Rate of direct I/O (for example, disk and tape) operations Buffered I/O Rate Rate of buffered I/O (for example, terminal and line printer) operations Mailbox Write Rate Rate of write-to-mailbox requests received by the system Split Transfer Rate Rate at which transfers were split into multiple I/Os Log Name Translation Rate Rate of logical name translations File Open Rate Rate at which files were opened Page Fault Rate Rate of occurrence of page faults for all working sets Page Read Rate Rate of pages read from disk as a result of page faults Page Read I/O Rate Rate of read I/O operations from disk as a result of page faults Page Write Rate Rate of pages written to the page file Page Write I/O Rate Rate of write I/O operations to the page file Inswap Rate Rate at which working sets were read into memory from the swap file Free List Size Number of pages on the free page list Modified List Size Number of pages on the modified page list
MONITOR> MONITOR /RECORD IO
|
This example shows that the direct I/O rate is currently at its highest level since the MONITOR command was entered and is significantly higher than the average rate. Termination of this command by Ctrl/C and entry of a MONITOR PROCESSES/TOPDIO command would show the top users of direct I/Os. Note that if I/O monitoring is begun at a later time, a new MONITOR request is defined. That is, it is not a continuation of the original request; the average, minimum, and maximum statistics are reinitialized. However, because the original request specified recording, that data can be played back for redisplay or summarization.
The MONITOR LOCK command initiates monitoring of the LOCK class.
MONITOR LOCK
/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
Specifies that a bar graph of average statistics is to be included in the display and summary outputs./CURRENT
Specifies that a bar graph of current statistics is to be included in the display and summary outputs. The /CURRENT qualifier is the default for the CLUSTER, MODES, STATES, SYSTEM, and VECTOR classes./MAXIMUM
Specifies that a bar graph of maximum statistics is to be included in the display and summary outputs./MINIMUM
Specifies that a bar graph of minimum statistics is to be included in the display and summary outputs.
The LOCK class includes the following data items:
Data Item Description New ENQ Rate Rate of new lock (ENQ) requests (as opposed to conversions) Converted ENQ Rate Rate of lock (ENQ) conversion requests DEQ Rate Rate of unlock (DEQ) requests Blocking AST Rate Rate of lock manager blocking ASTs delivered ENQs Forced To Wait Rate Rate of occurrence of locks that could not be granted immediately, thus causing a wait ENQs Not Queued Rate Rate of occurrence of locks that could not be granted immediately but requested not to be queued, and thus received an error status instead Deadlock Search Rate Rate at which a deadlock search was performed Deadlock Find Rate Rate at which deadlocks were found Total Locks Total number of locks in the system Total Resources Total number of resources in the system
MONITOR> MONITOR /RECORD IO MONITOR> MONITOR /INPUT=LOCKSTATS.DAT/SUMMARY/NODISPLAY LOCK/AVERAGE . . . MONITOR> [Ctrl/Z] $ TYPE MONITOR.SUM
|
This example shows the average use of the lock management subsystem during a typical workday, based on data that was previously recorded.
The MONITOR MODES command initiates monitoring of the MODES class, which includes a data item for each mode of processor operation.
MONITOR MODES
/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
Specifies that a bar graph of average statistics is to be included in the display and summary outputs./CPU
/NOCPU [=(x[,...])] (default)
In multiprocessor configurations, selects the CPU-specific form of output, where x specifies the CPU identification. If you specify /CPU without specifying a CPU identification, MONITOR displays MODES class statistics for each successive CPU until information for all active CPUs has been displayed. MONITOR then repeats the cycle beginning with the first CPU. If you specify one CPU identification, MONITOR displays statistics for that CPU only. If you specify multiple CPU identifications, MONITOR displays statistics for each successive CPU specified, then repeats the cycle beginning with the first specified CPU.Note that if you specify multiple CPU identifications, MONITOR does not notify you if one or more of the specified CPUs is unavailable. If all of the CPU identifications that you specify do not exist, then MONITOR will behave as if /CPU were specified without any arguments.
For multiprocessor systems, /NOCPU produces a single modes screen that reflects the combined time that all CPUs spent in each mode.
For nonmultiprocessor systems, the /CPU qualifier displays the CPU ID; /NOCPU does not display the CPU ID.
/CURRENT
Specifies that a bar graph of current statistics is to be included in the display and summary outputs. The /CURRENT qualifier is the default for the CLUSTER, MODES, STATES, SYSTEM, and VECTOR classes./MAXIMUM
Specifies that a bar graph of maximum statistics is to be included in the display and summary outputs./MINIMUM
Specifies that a bar graph of minimum statistics is to be included in the display and summary outputs./PERCENT
/NOPERCENT (default)
Controls whether statistics are expressed as percent values in display and summary outputs. The [NO]PERCENT qualifier is applicable only to the DISK, MODES, SCS, and STATES classes.
The following data items, included in the MODES class, can be displayed as percentages of all processor (CPU) time or as rates of clock ticks (10 millisecond units) per second:
Data Item Description Interrupt Stack (Interrupt State on Alpha and I64 systems) Time spent on the interrupt stack (on VAX systems) or in an interrupt state on a kernel stack (on Alpha and I64 systems). MP Synchronization Time spent synchronizing multiple CPUs (applicable to multiprocessor systems only). Kernel Mode Time spent in kernel mode, but not in an interrupt state. Executive Mode Time spent in executive mode. Supervisor Mode Time spent in supervisor mode. User Mode Time spent in user mode executing instructions. Compatibility Mode Time spent executing compatibility mode instructions. (This data item is meaningful only for VAX systems.) Idle Time Time not spent in any of the other modes. For multiprocessor systems, when you enter the MONITOR MODES command without using the /CPU qualifier to select specific CPUs, MONITOR produces a single modes screen similar to those produced for nonmultiprocessor systems. However, the statistics produced for multiprocessor systems reflect the combined time that all CPUs spent in each mode.
| #1 | |
|---|---|
MONITOR> MONITOR MODES /PERCENT
|
This display shows that, over the last collection interval, the processor spent 72 percent of its time executing user code, 8 percent executing system code to service user requests in executive and kernel modes, and 4 percent processing interrupts on the interrupt stack. It was idle 16 percent of the time. Time spent executing OpenVMS RMS code is included in executive-mode time. Time spent executing DCL code is included in supervisor-mode time.
If you omit the /PERCENT qualifier or specify /NOPERCENT, MONITOR displays mode times as rates of clock ticks per second, where a clock tick is 10 milliseconds. On a uniprocessor, the rate value is equivalent to the percent value.
| #2 | |
|---|---|
MONITOR> MONITOR MODES
|
This example demonstrates output of the MONITOR MODES command for a multiprocessor system. Displayed statistics represent rates of clock ticks per second. Information in the upper left corner of the screen indicates that node SAMPLE has four CPUs, two of which are active. Because the command line does not include the /CPU qualifier, statistics reflect the combined time that all CPUs spent in each mode.
The MONITOR MSCP_SERVER command initiates monitoring of the mass storage control protocol (MSCP) server class.
MONITOR MSCP_SERVER
/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
Specifies that a bar graph of average statistics is to be included in the display and summary outputs./CURRENT
Specifies that a bar graph of current statistics is to be included in the display and summary outputs. The /CURRENT qualifier is the default for the CLUSTER, MODES, STATES, SYSTEM, and VECTOR classes./MAXIMUM
Specifies that a bar graph of maximum statistics is to be included in the display and summary outputs./MINIMUM
Specifies that a bar graph of minimum statistics is to be included in the display and summary outputs.
The MSCP server class consists of the following data items that can be useful in tuning the MSCP server:
Data Item Description Server I/O Request Rate The rate at which remote processors request I/O transfers. Read Request Rate The rate at which remote processors request Read I/O transfers. Write Request Rate The rate at which remote processors request Write I/O transfers. Extra Fragment Rate The rate at which the server issues extra fragments. One or more extra fragments are created when, due to buffering constraints, the MSCP server issues multiple I/Os in order to fulfill a single I/O request. For example, if the MSCP server breaks up a 64-block request into 4 fragments of 16 blocks, 3 extra fragments are created. Fragmented Request Rate The rate at which fragmented requests occur. A fragmented request is a transfer request that the server fragments due to buffering constraints. For example, one fragmented request occurs when the server splits a 36-block request into 3 fragments of 16 blocks, 16 blocks, and 4 blocks. In this example, the server creates two extra fragments. Buffer Wait Rate The rate at which "buffer waits" occur in the server. A buffer wait occurs when a request must wait for MSCP buffer memory. Request Size Rates A histogram that displays the rate of requests for various block sizes.
MONITOR> MONITOR MSCP_SERVER
|
This example demonstrates use of the MONITOR MSCP_SERVER command to generate MSCP statistics on node GLOBBO.
| Previous | Next | Contents | Index |