Appendix E.1 System statistics

The following table shows the details of the system's statistics.

Table E-1 Details of the system's statistics

Types of statisticsEvent (obtained value)DetailsID for edition by dcreport#
RPC informationRPC call (response time)Length of time between sending a request from the client to the server in dc_rpc_call and dc_rpc_call_to, and receiving a response from the server.
This value is obtained when dc_rpc_call and dc_rpc_call_to are successful.
Synchronous-response RPC
Length of time between starting a dc_rpc_call or dc_rpc_call_to request, and receiving a response. The time includes the length of time for retries if the transmission of a request to the server in the dc_rpc_call or dc_rpc_call_to function fails and transmission is retried.
Chained RPC
Length of time between starting a dc_rpc_call or dc_rpc_call_to request, and receiving a response.
<Number of occurrences>
Number of times dc_rpc_call and dc_rpc_call_to are issued
1
User service execution (execution time)Length of time from the start of execution of a service function requested by dc_rpc_call and dc_rpc_call_to, until the service function issues a return value.
This value is obtained in dc_rpc_mainloop.
<Number of occurrences>
Number of times dc_rpc_call and dc_rpc_call_to are issued
2
RPC timeoutTimeout errors during RPC response wait processing.
<Number of occurrences>
Timeout error count
3
RPC failuresInternal failures during RPC processing.
<Number of occurrences>
Number of above failures
4
Schedule informationSchedule waits (queue length)The number of remaining service requests in the schedule queue of the user server (SPP).
When the queue length is lengthy, increase the number of resident processes specified by the parallel_count clause in the user service definition. If there is any nonresident process, decrease the specification of the balance_count clause to make it easier to start the nonresident process.
When the queue length is small, decrease the number of resident processes specified by the parallel_count clause in the user service definition, if necessary. If there is any nonresident process, increase the specification of the balance_count clause to make it harder to start the nonresident process. When there is difference between the average and the minimum value of the queue length, adjust the ratio of resident processes to nonresident processes specified in the balance_count clause, as well as the maximum number of processes.
<Number of occurrences>
Number of times the client made service requests (dc_rpc_call()) to the user server
11
Schedule (message size)The length of the service request message received by the user server (SPP).
The message length includes 512 bytes of control information added by the system.
It is preferable to specify a maximum message size in the message_buflen clause in the user service definition.
The following shows the range of preferable specifications for the message_store_buflen clause in the user service definition:
Message size (average) x queue length (max.) [Figure] message_store_buflen[Figure] message size (max.) x queue length (max.)
<Number of occurrences>
Number of times the client made service requests (dc_rpc_call) to the user server.
12
Size of the currently used buffer pool for storing messagesThe length of the buffer pool currently used by the user server (SPP) for storing messages, or shared by the scheduling buffer group, i.e., a group of user servers, to which the SPP belongs.
<Number of occurrences>
Number of times the client made service requests (dc_rpc_call) to the user server
13
Message size that could not be scheduled due to an insufficient buffer pool for storing messagesOf the service requests to the user server (SPP), the length of the service request messages that could not be scheduled due to an insufficient buffer pool for storing messages.
<Number of occurrences>
Number of times the user server failed to schedule service requests (dc_rpc_call) from the client due to an insufficient buffer pool for storing messages.
14
Lock informationLock acquisition (wait time)The time from the beginning of a lock wait status until the wait status is released.
The lock here means an internal lock by dc_lck_get, or an internal lock of the lock service issued by DAM, TAM, and MQA.
<Number of occurrences>
Number of times the lock wait occurred
21
Lock wait (queue length)Queue length when lock wait occurred.
<Number of occurrences>
Number of times the lock wait occurred
22
DeadlockNumber of deadlocks that occurred.
<Number of occurrences>
Same as above
23
DAM informationRead (input length)Input length of the DAM file data in the unit used for issuing dc_dam_read.
<Number of occurrences>
Number of times the dc_dam_read was issued
31
Read errorsErrors that occurred in the dc_dam_read interface part with the OS. Errors such as an invalid API argument are not counted.
<Number of time occurrences>
Count of above error
32
Write (output length)Output length of the DAM file data in the unit used for issuing dc_dam_write or dc_dam_rewrite.
<Number of occurrences>
Number of times the dc_dam_write or dc_dam_rewrite was issued
33
Write errorsOutput errors during actual input/output operations with a disk. Errors such as an invalid API argument are not counted.
<Number of occurrences>
Count of above error
34
Use of update buffer (update size)Amount of recovery journal (FJ)
Tune the size and number of journal files by calculating the amount of a journal for using DAM.
<Number of occurrences>
Number of times the recovery journal (FJ) was obtained.
35
The first DAM API issued in a transaction branch (total number of transaction branches)The number of transactions using DAM. The maximum value of this information is used as the peak number of transactions using DAM. To obtain an optimum memory (amount of dynamic shared memory for TP1/Server Base), specify the maximum value of this information for dam_tran_process_count in the DAM service definition.
<Number of occurrences>
Not collected.
36
Request for allocating a shared memory cache blockNumber of times the DAM file data block that is allocated in the resource manager's shared memory pool was allocated.
<Number of occurrences>
Number of events
38
Shared memory allocation (shared memory pool usage ratio)The periodically-obtained usage ratio of the resource manager shared memory pool.
We recommend that you specify the size of the shared memory for the resource manager (dam_cache_size in the DAM service definition) based on this value.
If the UAP frequently drops because the maximum value approaches 100%, increase the value of dam_cache_size.
<Number of occurrences>
Number of transactions that updated the DAM file specified with deferred update divided by 10
39
Process informationUAP abnormal terminationAbnormal terminations of UAP processes.
<Number of occurrences>
Sum of the UAP processes which terminated abnormally
41
Abnormal termination of system serverAbnormal terminations of system service processes.
<Number of occurrences>
Sum of the system servers which terminated abnormally
42
Process generation (total number of processes)Sum of the number of system service processes and UAP processes started in the OpenTP1 system at specified intervals.
<Number of occurrences>
Not collected
43
TAM informationTAM file real updating (number of written bytes)Number of data bytes (total value) actually updated in the TAM file during online operation.
To calculate the average number of bytes updated for one occurrence, use the following formula:
  • Average number of bytes updated for one occurrence = (current-maximum-value - previous-maximum-value) / number-of-occurrences
If there is no previous data, use the following formula:
  • Average number of bytes updated for one occurrence = (maximum-value - minimum-value) / (number-of-occurrences - 1)
<Number of occurrences>
Total number of data updates for the TAM file. Timer-start, cpd, or tamhold triggers the real update of the TAM file.
51
Commits, rollbacks (record reference count)The number of records referenced by one transaction. Since the TAM server outputs this value as statistics during commit processing, the number of occurrences is counted up no matter whether or not the dc_tam_read is issued. Therefore, a system that does not issue the dc_tam_read counts up the number of occurrences while the maximum value, minimum value and average remain zero.
<Number of occurrences>
Number of commits or rollbacks for TAM.
55
Commits, rollbacks (record update count)The number of records updated by one transaction. Since the TAM server outputs this value as statistics during commit processing, the number of occurrences is counted up no matter whether or not the dc_tam_write is issued. Therefore, a system that does not issue the dc_tam_write counts up the number of occurrences while the maximum value, minimum value and average remain zero.
<Number of occurrences>
Number of commits or rollbacks for TAM.
56
Transaction informationCommitsTransactions determined to be committed.
<Number of occurrences>
The number of transactions determined to be committed
61
RollbacksTransactions determined to be rolled back.
<Number of occurrences>
The number of transactions determined to be rolled back
62
Name informationCache hitsThe number of times the service information set in the service information cache area is referenced at the node.
The service information cache area means the area for setting service information of the server started at a different node as for the name service. Service information means the address information needed for RPC with the server. This information is set when a client requests reference to service information while the name service of the same node as for the client does not have the service information.
<Number of occurrences>
Number of times above information was referenced
71
Local hitsThe number of times the service information set in the service information local area is referenced in the node.
The service information local area means the area for setting service information of the server started at the same node as for the name service. Starting the server sets the service information.
<Number of occurrences>
Number of times above information was referenced
72
LookupThe number of times the reference to service information is requested in the node.
This data is counted up even if neither the service information local area nor the service information cache area stores service information.
<Number of occurrences>
Number of times the reference to service information is requested
73
Checkpoint dump informationCheckpoint dump collection (collection interval)Checkpoint trigger interval. This means the time interval between a previous checkpoint and the current checkpoint. Since a checkpoint occurs depending on output count of a regular journal block, the frequency is evaluated by time.
<Number of occurrences>
Number of checkpoint dumps triggered
81
Checkpoint dump validation (collection time)The time between triggering and validating the checkpoint dump collection. This means the time that elapsed before completing validation of checkpoint dump after an event triggers checkpoint dump and each system server starts dump processing.
<Number of occurrences>
Number of checkpoint dumps triggered
82
Message queue informationRead message (input length)The message length read from the queue file.
<Number of occurrences>
Read message length
91
Write message (output length)The message length written to the queue file.
<Number of occurrences>
Written message length
92
Read errorsFIL server errors during input from the physical file (abnormal cases).
<Number of occurrences>
Count of above error
93
Write errorsFIL server errors during output to the physical file (abnormal cases).
<Number of occurrences>
Count of above error
94
Waits for an available bufferThe number of times the necessary input/output buffers became insufficient due to too many requests concentrated on the queue file. Use this value to check the number of input/output buffers in the queue service definition.
<Number of occurrences>
Number of events
95
Real readsInput count from the physical file when there is no message in the input/output buffer. To reduce this value, increase the number of input/output buffers in the queue service definition.
<Number of occurrences>
Count of above input
96
Real writesOutput count to the physical file.
Compare this value with the number of write messages shown above to see how many write requests were input and/or output in batches.
To reduce the input/output count, increase the number of input/output buffers in the queue service definition.
The output count to the physical file should be zero when the deferred write facility is in use, because any output to the physical file is counted as the deferred writing (real write count).
<Number of occurrences>
Count of above output
97
Deferred writing (real write count)Output count to the physical file by deferred writing when the deferred write facility is being used.
<Number of occurrences>
Count of above output
99
Deferred writing for each physical file (number of records written in deferred mode)The value obtained for tuning the number of input/output buffers (quegrp -n option) in the message queue service definition.
<Number of occurrences>
Number of records written in deferred mode
151
Deferred writing for each physical file (messages written in deferred mode)The value obtained for tuning the number of input/output buffers (quegrp -n option) in the message queue service definition.
<Number of occurrences>
Percentage of messages written in deferred mode
152
Journal informationBuffer fullThe number of journal buffers for a system is twice of the number of parallel disk accesses, plus 1. One of these buffers is the current buffer.
Buffer full means the condition that an attempt to buffer the journal record to the current buffer failed due to insufficient buffer area available.
<Number of occurrences>
Number of events
102
Waits for available bufferThe condition of journal records waiting to be buffered because all journal buffers are full or in the process of being output.
<Number of occurrences>
Number of events
103
Journal output (block length)Data length of a journal block
Use this value to obtain the amount of the journal that occurred in the specified unit of time.
<Number of occurrences>
Journal block output count
104
Journal output (data length of the non-bus section)OpenTP1 internal information.
<Number of occurrences>
Number of events
105
Input/output waits (number of waiting buffers)The number of buffers waiting for output upon completion of journal output. The actual value times 100 is obtained to prevent the average value from being rounded.
<Number of occurrences>
Completed journal output count
107
WriteSystem's internal output count that includes both the journal data and file control information during a journal swap.
<Number of occurrences>
Count of above output
108
Write errorsErrors when data is written to a journal file.
<Number of occurrences>
Count of above error
109
Swap (swap time)The overhead time when journal files are swapped.
<Number of occurrences>
Journal file was swapping count
110
Journal input (data length)Length of data input from a journal file.
<Number of occurrences>
Count of data input from the journal file
111
ReadThe count for system's internal input that includes both the journal data and file control information during journal swap.
<Number of occurrences>
Above input count
113
Read errorsErrors when data is read from a journal file.
<Number of occurrences>
Count of above error
114
MCF informationMessage input to the memory queueWriting of messages to the memory queue on the OpenTP1 node.
<Number of occurrences>
Message write count shown above
121
Message input to the disk queueWriting of messages to the disk queue on the OpenTP1 node.
<Number of occurrences>
Message write count shown above
122
Message output from the memory queueRetrieving messages from the memory queue in the OpenTP1 node.
<Number of occurrences>
Message retrieval count shown above
123
Message output from the disk queueRetrieving messages from the disk queue in the OpenTP1 node.
<Number of occurrences>
Message retrieval count shown above
124
Disconnects due to a connection errorConnection errors in the OpenTP1 node.
<Number of occurrences>
Count of above error
125
Shared memory control informationSize of static shared memory usedTotal size of static shared memory blocks being used after static shared memory blocks have been allocated or released.
<Number of occurrences>
Number of statistics journals collected during the editing period + number of times the function for allocating or releasing static shared memory was issued
131
Maximum size of the required amount of static shared memoryMaximum size of the static shared memory blocks in use. This information is obtained after the allocation or release of blocks if the previous pool size has been updated. The pool size is the total size of the shared memory in use, and includes empty areas in the pool created by releasing allocated shared memory.
Accordingly, specify a value larger than this size in static_shmpool_size of the system environment definition.
<Number of occurrences>
Same as the size of static shared memory in use (if the previous pool size has been updated)
132
Size of dynamic shared memory usedTotal size of the dynamic shared memory blocks being used after dynamic shared memory blocks have been allocated or released.
<Number of occurrences>
Number of statistics journals collected during the editing period + number of times the function for allocating or releasing dynamic shared memory was issued
133
Maximum size of the required amount of dynamic shared memoryMaximum size of the dynamic shared memory blocks in use. This information is obtained after allocation or release of blocks if the previous pool size has been updated. Specify a value larger than this size in dynamic_shmpool_size of the system environment definition.
<Number of occurrences>
Same as the size of dynamic shared memory in use (if the previous pool size has been updated)
134
MQA service informationGet message (input length)The length of the message that was read from the queue by the application and MQT server.
<Number of occurrences>
Number of times the AP and MQT server retrieved messages from the queue
161
Put message (output length)The length of the message that was entered into the queue by the application and MQT server.
<Number of occurrences>
Number of times the AP and MQT server entered messages to the queue
162
Read errorsI/O errors during input of messages from a file.
<Number of occurrences>
Count of above error
163
Write errorsI/O errors during output of messages to a file.
<Number of occurrences>
Count of above error
164
Waits for an available bufferThe number of times the buffers became insufficient.
Increase the number of input/output buffers if this value is counted.
<Number of occurrences>
Number of events
165
Real readsI/O for reading messages from the file.
If this value is counted, increase the number of input/output buffers to reduce I/O for receiving messages.
<Number of occurrences>
Number of events
166
Real writesI/Os for writing messages to a file.
<Number of occurrences>
Number of events
167
Write request timeProcessing time for writing messages to a file (I/O time + time for processing of request to the MQAI/O process).
<Number of occurrences>
Number of requests for writing messages to the file
169
Number of messages activated at one triggerNumber of messages left in the queue when a trigger was generated.
<Number of occurrences>
Trigger message generation count
170
Wait time for starting message transferThe time between issuing MQPUT and startup of message transfer.
<Number of occurrences>
Number of messages transferred
171
Wait time for message arrivalThe time between issuing MQPUT and completion of message transfer.
<Number of occurrences>
Number of messages transferred
172
IST service informationInput lengthInformation about the input length of the IST table data in the unit used for issuing dc_ist_read.
<Number of occurrences>
Number of times the dc_ist_read was issued
191
Output lengthInformation about the output length of the IST table data in the unit used for issuing dc_ist_write.
<Number of occurrences>
Number of times the dc_ist_write was issued
192
Number of updates from other nodesNumber of update messages received prompting the local node to apply the IST table update operations performed at other nodes.
<Number of occurrences>
Number of events
193
Number of updates to other nodesNumber of times the local node updated the IST table and notified other nodes of the updating.
<Number of occurrences>
Number of events
194
XATMI service informationXATMI call (response time)The time between sending a service request from the client and receiving a response during OSI TP communication using the XATMI interface.
Synchronous-response
The time between sending a service request with tpcall and receiving a response.
Asynchronous-response
The time between sending a service request with tpacall and receiving a response with tpgetrply.
Non-response
Invalid value.
<Number of occurrences>
Number of times the tpcall or tpacall was issued. Those that returned by error before sending a service request are not counted.
211
XATMI user service execution timeThe time from the beginning of execution of a service function at the server during OSI TP communication using the XATMI interface until the service function issues return.
<Number of occurrences>
Number of times the user service was executed.
212
XATMI service failure countNumber of failures in OSI TP communication using the XATMI interface
Client side
Failures that occurred within a transaction after the service request was sent but before the transaction is determined, and failures that occurred outside a transaction after the service request was sent but before the client receives a response from the service. These failures include communication failures, server UAP failures, and error responses from the service.
Server side
Failures that occurred within a transaction after the service request was received but before the transaction is determined, and failures that occurred outside a transaction after the service request was received but before the service function terminates.
These failures include communication failures, client UAP failures, and rollback directions from the client.
<Number of occurrences>
Number of above failures.
213
#
Argument to be specified when the dcreport command is executed.

The following table lists the items to be edited in the system statistics.

The Unit in the table is the number of times an event occurred or the unit of the obtained value if there is one. If a particular event did not occur (the number of occurrences is 0), the edited values are meaningless.

Table E-2 Items to be edited in system statistics

Statistics typeEdited contentsUnitEditing in units of user serversID for edition by dcreport#1
Event (obtained value)NumberEdited values
AvgMaxMin
RPC informationRPC calls (response times)YYYYMicro- secondsY1
Execution of user service (execution times)YYYYMicro- secondsY2
RPC overtimesYNNNNumber of eventsY3
RPC errorsYNNNNumber of eventsY4
Scheduling informationSchedule wait (wait queue length)YYYYRequestsY11
Schedule (message size)YYYYBytesY12
Size of the currently used buffer pool for storing messagesYYYYBytesY13
Message size that could not be scheduled due to insufficient buffer pool for storing messagesYYYYBytesY14
Lock informationLock acquisition (wait time)YYYYMilli- secondsN21
Lock wait (wait queue length)YYYYRequestsN22
DeadlocksYNNNNumber of eventsN23
DAM informationread (input length)YYYYBytesY31
read errorsYNNNNumber of eventsY32
write (output length)YYYYBytesY33
write errorsYNNNNumber of eventsY34
Use of update buffer (update size)YYYYBytesY35
First DAM API issued in a transaction branch (total number of transaction branches)#2YYYTransaction branchesY36
Shared memory cache block allocation requestYNNNNumber of eventsY38
Shared memory allocation (shared memory utilization rate)#3YYYY%N39
Process informationUAP abnormal terminationsYNNNNumber of eventsN41
System server abnormal terminationsYNNNNumber of eventsN42
Process generations (total number of processes)#2YYYProcessesN43
TAM informationTAM file real updating (number of written bytes)YYYYBytesN51
Commits, rollbacks (number of record referencing operations)YYYYNumber of eventsY55
Commits, rollbacks (number of record updating operations)YYYYNumber of eventsY56
Transaction informationCommitsYNNNNumber of eventsY61
RollbacksYNNNNumber of eventsY62
Name informationCache hitsYNNNNumber of eventsY71
Local hitsYNNNNumber of eventsY72
lookupYNNNNumber of eventsY73
Checkpoint dump informationCollection of checkpoint dump (collection interval)YYYYMilli- secondsN81
Validation of checkpoint dump (collection time)YYYYMilli- secondsN82
Message queue informationRead message (input length)YYYYBytesN91
Write message (output length)YYYYBytesN92
Read errorsYNNNNumber of eventsN93
Write errorsYNNNNumber of eventsN94
Waits for an available bufferYNNNNumber of eventsN95
Real readsYNNNNumber of eventsN96
Real writesYNNNNumber of eventsN97
Deferred writing (real write count)YYYYTimesN99
Deferred writing for each physical file (number of records that were written in deferred mode)YYYYNumber of recordsN151
Deferred writing for each physical file (percentage of messages that were written in deferred mode)YYYY%N152
Journal informationBuffer full occurrencesYNNNNumber of eventsY102
Waits for an available bufferYNNNNumber of eventsY103
Journal output (block length)YYYYBytesN104
Journal output (data length of non-bus section)#4YYYYBytesN105
Input/output waits (Number of waiting buffer faces)YYYYNumber of eventsN107
writeYNNNNumber of eventsN108
Write errorsYNNNNumber of eventsN109
Swap (swap time)YYYYMicro- secondsN110
Journal input (data length)YYYYBytesN111
readYNNNNumber of eventsN113
Read errorsYNNNNumber of eventsN114
MCF informationMessage inputs to memory queueYNNNNumber of eventsN121
Message inputs to disk queueYNNNNumber of eventsN122
Message outputs from memory queueYNNNNumber of eventsN123
Message outputs from disk queueYNNNNumber of eventsN124
Disconnects due to a connection errorYNNNNumber of eventsN125
Shared memory control informationSize of static shared memory usedYYYYBytesY131
Maximum static shared memory size requirementYYYYBytesY132
Size of dynamic shared memory usedYYYYBytesY133
Maximum dynamic shared memory size requirementYYYYBytesY134
MQA service informationGet message (input length)YYYYBytesY161
Put message (output length)YYYYBytesY162
Read errorsY------Number of eventsY163
Write errorsY------Number of eventsY164
Waits for an available bufferY------Number of eventsY165
Real readsY------Number of eventsY166
Real writesY------Number of eventsY167
Write request timeYYYYMicro- secondsY169
Number of messages activated at one triggerYYYYNumber of eventsY170
Time of wait for message transfer startYYYYMilli- secondsY171
Time of wait for message arrivalYYYYMilli- secondsY172
IST service informationInput lengthYYYYBytesY191
Output lengthYYYYBytesY192
Number of updates from other nodesYNNNNumber of eventsN193
Number of updates to other nodesYNNNNumber of eventsN194
XATMI service informationXATMI call response timeYYYYMicro- secondsY211
XATMI user service execution timeYYYYMicro- secondsY212
XATMI service error countYNNNNumber of eventsY213
Legend:
Number: Number of occurrences
Avg: Average
Max: Maximum
Min: Minimum
Y: Can be output to a system journal file or can be edited in units of user servers.
N: Cannot be output to a system journal file or cannot be edited in units of user servers.
--: Neither edited nor output.
#1
Argument to be specified when the dcreport command is executed.
#2
Output when the dcreport command is executed.
#3
These statistics are not output if you do not update the DAM file with the deferred update specification. Note that these statistics are output once for every 10 deferred updates.
#4
This is OpenTP1 internal information.