For Windows, use lowercase alphabetic characters to specify operands.
For details about the operands that are specific to TP1/Client/W, see TP1/Client/W-specific operands. For details about the operands that are specific to TP1/Client/P, see TP1/Client/P-specific operands.
- DCNAMPORT=name-service-port-number ~<unsigned integer> ((5001-65535)) <<10000>>
Specify the port number of the name service. The target TP1/Server must use this port number to start the name service.
You can also specify the port number of the name service using DCHOST in the client environment definition.
- DCHOST=TP1/Server-as-a-gateway ~<character string>
Specify the host computer name and the port number for the TP1/Server used as a gateway. You can specify two or more TP1/Servers as gateways using a comma (,) as a separator.
- Format:
- host-computer-name[:port-number][,host-computer-name[:port-number],...]
- host-computer-name~<character string>
- port-number~<unsigned integer>((5001-65535))
- You can specify a maximum of 63 characters for the host name. When 00000008 is specified for DCCLTOPTION in the client environment definition, you can specify a maximum of 255 characters. The maximum number of characters you can specify in this operand is 1,023.
- Do not place a blank character (space or tab) except after the separator (,).
You can specify an IP address in decimal dot notation for the host name.
When the port number is not specified, the system uses the value for DCNAMPORT in the client environment definition.
When you have specified more than one TP1/Server in DCHOST and an error is detected in the TP1/Server being used as a gateway, system operation depends on the specification of DCHOSTSELECT in the client environment definition. If N is specified for DCHOSTSELECT, the system attempts to replace the failed node by referencing the next TP1/Server of the currently used TP1/Server. If Y is specified for DCHOSTSELECT, the system selects a TP1/Server at random (excluding the TP1/Server in which the error was detected) and attempts to replace the failed node.
Note that when you issue a client user authentication request with the gateway TP1/Server specified, this specification overrides the specification of DCHOST in the client environment definition. If the gateway TP1/Server is not specified either in a client user authentication request or in DCHOST in the client environment definition, TP1/Client uses a broadcast to determine the TP1/Server to be used as the gateway. To perform a broadcast in TP1/Client/P, you must specify the broadcast address in the hosts file (the host name must be broadcast).
- DCWATCHTIM=maximum-response-wait-time~<unsigned integer> ((0-65535)) <<180>> (unit: seconds)
For response type RPCs, specify the maximum wait time between a CUP sending a service request to an SPP and the return of the service response. An error is returned to the CUP if no response is received within the specified duration.
If DCWATCHTIM is set to zero, the function waits for a response indefinitely.
- DCCLTCONNECTTIMEOUT=timeout-for-establishing-connection ~<unsigned integer>((0-65535))<<0>>(unit: seconds)
Specify the timeout for establishing connection in non-blocking mode for data transmission.
When you specify 0 or omit this operand, connection is established in blocking mode and the operating system monitors whether connection is established.
If this operand is incorrectly specified, the KFCA02401-E message is output to the error log and the dc_clt_cltin_s function returns a DCCLTER_FATAL error.
The function issued by the CUP may return an error before the time specified in this operand elapses. This error occurs, for example, when the remote system is not running. This error occurs because the operating system's timeout for establishing a connection prevails over the timeout specified in this operand. The operating system's timeout for establishing a connection differs depending on the platform you are using.
This operand monitors the time elapses after a connection establishment function of TP1/Client is executed. This time does not include the processing time of the function itself. Depending on the function or program you use, its processing time may be longer than the time specified in this operand.
- DCCLTTREXPTM=expiry-time-in-transaction-branch ~<unsigned integer> ((0-65535)) (Unit: seconds)
Specify the maximum expiry time in a transaction branch. This operand is valid only when the transaction is started from a CUP.
If the transaction branch is not completed within the specified expiry time, its process is terminated abnormally and rolled back. Specify 0 not to perform monitoring.
If this operand is not specified, the process follows the specification of the trn_expiration_time operand in the client service definition. If the connection destination is a RAP-processing server, the process follows the specification of the trn_expiration_time operand in the RAP-processing listener service definition.
When using the RPC function, use DCCLTTREXPSP to specify whether the processing time for transaction branches executed in other processes is to be included in the monitoring time.
- DCCLTTREXPSP=Y|N|F
Specify whether the following time is to be included in the monitoring time for a transaction branch:
The time required for a transaction branch of the transactional RPC executing process to call another transaction branch using the RPC function and to wait for completion of the processing
Y: Includes the time in the monitoring time.
N or F: Does not include the time in the monitoring time.
If this operand is not specified, the process follows the specification of the trn_expiration_time_suspend operand in the client service definition. If the connection destination is a RAP-processing server, the process follows the specification of the trn_expiration_time_suspend operand in the RAP-processing listener service definition.
- DCCLTTRWATTM=maximum-time-interval-in-transaction-inquiry-response~ <unsigned integer> ((1-65535))<<180>>(Unit: seconds)
Specify the maximum time interval in transaction processing between an inquiry made from the CUP to a server (by issuing a transaction control function or the dc_rpc_call_s function) and another inquiry. This operand is valid only when a transaction is started from the CUP.
The maximum time interval in transaction inquiry response must be smaller than the expiry time in transaction branch.
If no inquiry occurs within the specified time, the system rolls back the transaction process on the server.
- DCCLTTRCPUTM=CPU-monitoring-time-in-transaction-branch ~<unsigned integer> ((0-65535))(Unit: seconds)
Specify the CPU time that can be used by a transaction branch before synchronous-point processing. This operand is valid only when the transaction is started from a CUP.
If 0 is specified, the CPU time is not monitored.
If the specified time is exceeded, the process of the transaction branch is terminated abnormally and rolled back.
If this operand is not specified, the process follows the specification of the trn_cpu_time operand in the client service definition. If the connection destination is a RAP-processing server, the process follows the specification of the trn_cpu_time operand in the RAP-processing listener service definition.
- DCCLTUTTRCMT=Y|N ~<<N>>
Specify whether the transaction started from a CUP using the online tester is to be committed or rolled back.
Y: The transaction is committed.
N: The transaction is rolled back.
- DCRCVPORT=Receive-CUP-port-number ~<unsigned integer>((1-65535)) <<11000>>
Specify the port number of the CUP that receives messages when you use the TCP/IP communication facility to receive messages. Specify this port number on the message-originating side. Specify a unique port number for each process or thread when multiple processes or multiple threads are executed simultaneously on the same machine.
Do not specify a port number for use by the operating system or other programs even if one can be specified. If you specify a port number in this case, response data might not be received correctly. The port numbers used by the operating system differ depending on the operating system. For details, see the documentation of your operating system.
- DCSNDHOST=connected-node-name ~<character string>
Specify the host name of the node to be connected by establishing a connection when you use the TCP/IP communication facility to send messages.
You can specify an IP address in decimal dot notation as a host name.
You can specify a maximum of 63 characters for the host name. You can specify a maximum of 255 characters when 00000008 is specified for DCCLTOPTION in the client environment definition.
- DCSNDPORT=connected-port-number ~<unsigned integer>((1-65535)) <<12000>>
Specify the port number of the node to be connected by establishing a connection when you use the TCP/IP communication facility to send messages.
- DCSOCKOPENATRCV=Y|N ~<<N>>
Specify when to open the receive socket for sending and reception performed with one connection, when using the TCP/IP communication facility. Here, when to open the receive socket means when TP1/Client begins to wait for connection from the other party.
This definition is effective only when DCCLT_SNDRCV is specified in the flags argument of the dc_rpc_open_s function.
Y: Opens the receive socket if no connection has been established when any of the following functions is issued:
- dc_clt_receive_s
- dc_clt_receive2_s
- dc_clt_assem_receive_s
N: Opens the receive socket when the dc_rpc_open_s function is executed. This is the default.
- DCCLTDELIVERYCHECK=Y|N ~<<N>>
Specify whether to use the message delivery confirmation facility.
Y: The message delivery confirmation facility is used.
N: The message delivery confirmation facility is not used.
When the message delivery confirmation facility is used, if the dc_clt_assem_send_s function is issued, the function sends a message, receives response-only data, and then returns control. If the dc_clt_assem_receive_s function is issued, the function receives a message, sends response-only data, and then returns control.
If either the dc_clt_assem_send_s or dc_clt_assem_receive_s function is issued when this operand is set to N, the function uses the message assembly facility to send or receive a message.
- DCUTOKEY=test-user-ID ~<1-4 alphanumeric characters>
Specify this operand when executing a CUP using the online tester.
With this operand specified, SPPs started from the CUP can be executed in test mode.
- DCCACHE=number-of-areas-for-temporarily-storing-service-information ~<unsigned integer> ((2-10240)) <<8>>
For when you execute an RPC from a client, specify the number of cache areas that store the service information provided by the TP1/Server name service that is used as a gateway. Each cache area stores one service information item.
The stored service information is deleted from the cache areas when the effective period, specified using DCCLTCACHETIM in the client environment definition, expires.
When you specify this operand, use the following as a guideline:
- When DCCLTLOADBALANCE=N is specified
- Specify the number of servers to which the client sends RPC-based requests.
- When DCCLTLOADBALANCE=Y is specified
- Specify the number of all servers in the nodes that are started as multi-node servers.
Each cache area consumes about 150 bytes of memory. When DCSCDDIRECT=Y is specified in the client environment definition, this operand is invalid.
- DCCLTCACHETIM=expiration-of-temporarily-stored-service-information ~<unsigned integer> ((0-65535)) <<30>> (Unit: seconds)
Specify the effective period for the service information that is acquired from the TP1/Server name service that is used as a gateway. When the effective period expires, the service information is deleted from the cache areas. When you specify 0, no effective period is specified. Once the service information is stored, it will be valid until the dc_rpc_close_s function is issued or the existing information is overwritten by new service information when free cache areas run short. This operand is valid only when DCCLTLOADBALANCE=Y is specified in the client environment definition. This operand is invalid when DCSCDDIRECT=Y is specified in the client environment definition.
- DCCLTLOADBALANCE=Y|N ~<<N>>
Specify whether to use the inter-node load-balancing facility when multi-node servers are used. The inter-node load-balancing facility evaluates the load status of each node internally within TP1/Client when an RPC is executed, and distributes the load to servers with less load.
Y: Uses the inter-node load-balancing facility.
N: Does not use the inter-node load-balancing facility.
When you specify Y for this operand, specify the following operands in the client environment definition according to the number of nodes or the number of servers that handle RPC-based requests.
- DCCACHE
- DCCLTCACHETIM
- DCCLTNAMEXTEND
This operand is invalid when DCSCDDIRECT=Y is specified in the client environment definition.
- DCCLTSERVICEGROUPLIST=file-defining-correspondence-between-service-groups-and-RPCs ~<character string>
Specify a text file (with its path name) that defines correspondence between the service group and the RPC entry point for the corresponding server. This file is used for issuing an RPC to a server other than OpenTP1. Define this file as follows.
Format:
<service-group-name> <server-host-computer-name> <entry-point-port-number> [,<server-host-computer-name> <entry-point-port-number>,...] [<comment>]
- Define a pair of the service group and the RPC entry point per line in a text file. When two or more RPC entry points are available, use as many lines as these entry points.
- Separate each item with a space or a tab.
- Specify each item as follows.
- service-group-name
- Any character string with up to 31 characters.
- server-host-computer-name
- Host name of the connection destination host. You can specify a maximum of 63 characters for the host name. You can specify a maximum of 255 characters when 00000008 is specified for DCCLTOPTION in the client environment definition. You can also specify an IP address in decimal dot notation as a host name.
- entry-point-port-number
- Numeric value between 1 and 65535 as a port number for accepting RPCs.
- comment
- Any text beginning with #. This is assumed to be a comment to the end of the line and is ignored for the processing. The comment is optional.
When you specify multiple RPC entry points, the system selects one of them at random and attempts to connect it. If an attempt to connect the selected host fails, that host is eliminated from the connection destination options. Then, the system selects another RPC entry point at random in an attempt to connect. This step will be repeated. If all the attempts to connect a RPC entry point fail, the dc_rpc_call_s function returns an error.
- Evaluating the file content:
- When the file is defined incorrectly, TP1/Client ignores the invalid line. It also outputs the corresponding line number to the error log.
If the name of the service group called by the dc_rpc_call_s function has not been defined in the file specified in this operand, operation follows the specification of the DCCLTNOSERVER operand in the client environment definition:
- When DCCLTNOSERVER=Y is specified:
The function immediately returns a DCRPCER_NO_SUCH_SERVICE_GROUP error.
- When DCCLTNOSERVER=N is specified or the DCCLTNOSERVER operand is omitted:
On recognizing that the name of the called service group is not defined in the file specified in this operand, the function performs an RPC to TP1/Server.
- DCCLTCONNECTRETRY=retry-count-for-connection-establishment ~<unsigned integer> ((0-255)) <<0>>
Specify the maximum number of times a request to establish connection should be retried if a timeout occurred for a request because the server was offline or turned off, for example. When 0 is specified in this operand or when this operand is not specified, TP1/Client does not perform a retry. If this operand is specified incorrectly, TP1/Client assumes 0. This definition is effective when you attempt user authentication (specify DCCLTAUTHENT=Y and execute the dc_clt_cltin_s function with DCNOFLAGS specified in the flags argument).
- DCSCDDIRECT=Y|N ~<<N>>
Specify whether to use the function to directly inquire into the schedule service without inquiring service information from the TP1/Server name service (RPC that does not use the name service).
Y: Uses an RPC that does not use the name service.
N: Does not use an RPC that does not use the name service.
When DCSCDPORT in the client environment definition specifies the port number for the schedule service, the client makes an inquiry using that port number. When no DCSCDPORT is defined, the client acquires the port number for the schedule service from TP1/Server, then makes an inquiry.
When this function is used, the program cannot call socket-receiving type SPPs. Definitions of DCCACHE, DCCLTCACHETIM, and DCCLTLOADBALANCE in the client environment definition are ignored. When DCCLTSERVICEGROUPLIST is specified in the client environment definition, the definition of DCSCDDIRECT is ignored.
- DCSCDPORT=schedule-service-port-number ~<unsigned integer> ((5001-65535))
Specify the port number of the schedule service. When DCSCDMULTI=Y and DCSCDDIRECT=Y are specified in the client environment definition, specify the port number which is used as the base of the multi-scheduler daemon. The target TP1/Server must use this port number to start the schedule service or the multi-scheduler daemon. For details about the specification of the schedule service or the multi-scheduler daemon of TP1/Server, see the manual OpenTP1 System Definition.
This operand is valid only when DCSCDDIRECT=Y is specified in the client environment definition.
When DCSCDMULTI=Y and DCSCDDIRECT=Y are specified in the client environment definition, also see the description of DCSCDMULTICOUNT in the client environment definition.
If you omit this operand, the name service is sent a request regarding the port number of the schedule service or the multi-scheduler daemon.
- DCCLTDATACOMP=Y|N ~<<N>>
Specify whether to use the data compression.
Y: The data compression is used.
N: The data compression is not used.
- DCEXTENDFUNCTION=facility-extension-level-of-RPC-service ~<unsigned hexadecimal integer> ((00000000-00000001)) <<00000000>>
Specify either of the following as the extension level of the RPC service facility. Specify all the zeros written below. Even if you specify a value that cannot be specified, TP1/Client may operate incorrectly without generating an error code.
- 00000000
- The RPC service facility is not extended.
- 00000001
- If the SPP that is executing a service request terminates abnormally, the dc_rpc_call_s function return an error code (DCRPCER_SERVICE_TERMINATED) to isolate the error. If this specification is not made, DCRPCER_TIMED_OUT and DCRPCER_SERVICE_NOT_UP are returned.
This definition is invalid when a service request (dc_rpc_call_s function) is executed during establishment of a permanent connection or within the scope of a transaction. During establishment of a permanent connection or within a transaction, the specification of the rpc_extend_function operand in the user service default definition is valid.
- DCCLTINQUIRETIME=maximum-time-interval-in-permanent-connection <unsigned integer> ((0-1048575)) (unit: seconds)
Specify the maximum interval between an inquiry from the CUP to the server and the next inquiry. The CUP execution process or RAP-processing-server monitors this interval, and forcibly releases the permanent connection if no inquiry is made within the specified period of time.
This definition is ineffective when you establish permanent connection with a DCCM3 logical terminal. DCCM3 provides the equivalent feature using the terminal monitoring time. For DCCM3, specify the terminal monitoring time in the LEFTLIMIT clause of the TERMINAL statement for the data communication definition.
If expiration of the maximum interval is detected in a transaction, the transaction is rolled back.
If 0 is specified, the system waits infinitely for an inquiry from the CUP. When you omit this specification, the process follows the specification of the clt_inquire_time operand in the client service definition or rap_inquire_time operand in the RAP-processing listener service definition. If the connection destination is a RAP-processing server, the process follows the specification of the rap_inquire_time operand in the RAP-processing listener service definition.
The maximum time interval in permanent connection specified in the definition monitors the time between the dc_clt_connect_s function and the dc_clt_disconnect_s function when a permanent connection is established. The maximum time interval in transaction inquiry response (DCCLTTRWATTM) monitors the time between inquiries up to the dc_trn_unchained_commit_s function when the dc_trn_begin_s function is issued without establishing a permanent connection.
- DCCLTPORT=client-extended-service-port-number ~ <unsigned integer> ((5001-65535))
Specify the port number of the client extended service. The target TP1/Server must use this port number to start the client extended service. Specify the port number of the client extended service using the clt_port operand in the client service definition.
If this definition is not specified, the port number of the client extended service is inquired from the name service.
- DCCLTDCCMHOST=DCCM3-logical-terminal-host-name
When you request establishing the permanent connection with a DCCM3 logical terminal, specify the host name of the connection target logical terminal. At this time, specify DCCLT_DCCM3 in the flags argument of the dc_clt_connect_s function.
Specify as follows.
- host-name:[port-number][,host-name[:port-number],...]
- host-name ~<character string>
- port-number ~<unsigned integer> ((1-65535))
- You can specify a maximum of 63 characters for the host name. When 00000008 is specified for DCCLTOPTION in the client environment definition, you can specify a maximum of 255 characters. The maximum number of characters you can specify in this operand is 1,023.
- You can use a blank character (space or tab) only after a delimiter (,).
You can specify an IP address in decimal dot notation as a host name.
If you do not specify a port number, the port number of the DCCM3 logical terminal specified in the DCCLTDCCMPORT client environment definition is assumed.
When you specify multiple DCCM3 logical terminals, the system selects one of them at random and attempts to connect it. If an attempt to connect to the selected DCCM3 logical terminal fails, that DCCM3 logical terminal is eliminated from the connection destination options. Then, the system selects another DCCM3 logical terminal at random in an attempt to connect. This step will be repeated. If all the attempts to connect a DCCM3 logical terminal fail, the dc_clt_connect_s function returns an error.
When you establish a permanent connection to communicate with a DCCM3 logical terminal, the value specified in the DCCLTSERVICEGROUPLIST client environment definition is ignored. Note that the data compression cannot be performed.
- DCCLTDCCMPORT=DCCM3-logical-terminal-port-number ~ <unsigned integer> ((1-65535)) <<30000>>
Specify the port number with which the CUP makes a request of permanent connection to the DCCM3 logical terminal.
- DCCLTXATMI=Y|N ~ <<N>>
Specify whether to use the XATMI interface for communication. If this operand is specified incorrectly, TP1/Client assumes N.
Y: Use the XATMI interface.
N: Do not use the XATMI interface.
- DCWATCHTIMINHERIT=Y|N ~ <<N>>
Specify whether an extended client service should inherit the maximum CUP response wait time for transaction or connection control.
Y: An extended client service inherits the maximum CUP response wait time.
N: An extended client service does not inherit the maximum CUP response wait time.
When you specify Y for this operand, refer to DCCLTDELAY in the client environment definition.
- DCCLTDELAY=maximum-communication-delay-time ~ <unsigned integer> ((0-65535)) <<0>> (units: seconds)
Specify this item when terminating server response monitoring earlier than the client response monitoring, considering the communication overhead between the CUP and an extended client service. This definition allows terminating the server monitoring quickly by the specified time. Prompt termination prevents the improper transmission of messages due to the time-out of client monitoring.
This definition is valid only when Y is set for DCWATCHTIMINHERIT in the client environment definition. The definition of DCCLTDELAY is ignored when 0 is set for client environment definition DCWATCHTIM. If the subtraction of a value specified in DCCLTDELAY from a value specified in DCWATCHTIM provides 0 or negative value, the definition of DCCLTDELAY is ignored, with 1 defaulted to.
You can dynamically change a value specified for DCWATCHTIM in the client environment definition by issuing the dc_rpc_set_watch_time_s function. If such value has been dynamically changed, the new value is calculated by the dc_rpc_watch_time_s function.
- DCCLTCUPSNDHOST=CUP-send-host ~<character string>
Specify the host that sends a connection establishment request.
You can specify a maximum of 63 characters for the host name. If you specify 00000008 for DCCLTOPTION in the client environment definition, you can specify a maximum of 255 characters for the host name.
You can also specify an IP address in decimal dot notation for the host name.
If you specify localhost as the host name or an IP address that begins with 127, the dc_clt_cltin_s function returns a DCCLTER_FATAL error.
If you specify a host that is not on the same machine on which the CUP runs, the communication function returns a DCCLTER_NET_DOWN or DCRPCER_NET_DOWN error.
If you omit this operand, the send host is assigned automatically.
- DCCLTCUPRCVPORT=port-number-used-for-CUP-reception ~ <unsigned integer> ((5001-65535))
Specify the port number of a CUP that receives a message from the server.
The port number specified in this definition is valid when you use the following functions.
- Ordinary RPC function (for reception)
- Transaction control function
- Permanent connection establishment function
If you omit this operand, the system uses the port number assigned by it.
Specify a unique port number for each process or thread when multiple processes or multiple threads are executed simultaneously on the same machine.
Do not specify a port number for use by the operating system or other programs even if the port number is valid. If you specify a port number in this case, response data might not be received correctly. The port numbers used by the operating system differ depending on the operating system. For details, see the documentation of your operating system.
- DCCLTRAPHOST=RAP-processing-listener or DCCM3-logical-terminal
Specify the host name and port number of a RAP-processing listener or the DCCM3-logical-terminal supported in TP1/Server.
The following items give the formats.
- host-name:port-number[,host-name:port-number,...]
- host-name ~ <character string>
- port-number ~ <unsigned integer> ((5001-65535))
- You can specify a maximum of 63 characters for the host name. When 00000008 is specified for DCCLTOPTION in the client environment definition, you can specify a maximum of 255 characters. The maximum number of characters you can specify in this operand is 1,023.
- You can use a blank character (space or tab) only after a delimiter (,).
You can also specify an IP address in decimal dot notation as a host name.
If you issue the dc_clt_connect_s function with DCNOFLAGS set at flags, the system requests a RAP-processing listener or the DCCM3 logical terminal for TP1/Server to establish a permanent connection when you specify this definition. If you do not specify this definition, the system requests an extended client service for TP1/Server to establish a permanent connection.
When you request establishing the permanent connection with a DCCM3 logical terminal, the data compression is unavailable.
Note, however, that, if a firewall exists between the CUP and the RAP-processing listener or the DCCM3 logical terminal, the host name and port number specified in DCCLTRAPHOST become the host name and port number of the firewall, respectively.
When you specify several pairs of a host name and the port number of the RAP-processing listener (or the port number of the DCCM3 logical terminal), the system selects one of the pairs at random and attempts to connect the selected host. If an attempt to connect the selected host fails, that host is eliminated from the connection destination options. Then, the system selects another host at random in an attempt to connect. This step will be repeated. If all the attempts to connect a host fail, the dc_clt_connect_s function returns an error.
When you establish a permanent connection to communicate with a RAP-processing listener or with a DCCM3 logical terminal, the value specified in the DCCLTSERVICEGROUPLIST client environment definition is ignored.
The following table shows the relationships between the client environment definition and flags on the dc_clt_connect_s function. You can find the target of permanent connection.
Argument flags |
Client environment definition |
Establishing permanent connection for: |
DCCLTDCCMHOST |
DCCLTRAPHOST |
DCNOFLAGS
|
Y
|
Y |
RAP-processing server or DCCM3 logical terminal1 |
|
CUP execution process |
|
Y |
RAP-processing server or DCCM3 logical terminal1 |
|
CUP execution process |
DCCLT_DCCM3
|
Y
|
Y |
DCCM3 logical terminal2 |
|
DCCM3 logical terminal2 |
|
Y |
Error return |
|
Error return |
- Legend:
- Y: Specified
- Blank: Not specified
- 1
- Establish permanent connection for a DCCM3 logical terminal specified by DCCLTRAPHOST.
- 2
- Establish permanent connection for a DCCM3 logical terminal specified by DCCLTDCCMHOST.
- DCCLTRAPAUTOCONNECT=Y|N ~ <<N>>
Specify whether to automatically establish permanent connection between a CUP and a RAP-processing server or between a CUP and a DCCM3 logical terminal.
Y: Automatically establishes a permanent connection.
N: Does not automatically establish a permanent connection.
Specifying Y in this definition automatically establishes the permanent connection if the permanent connection is not yet established when the following functions are executed. The requested permanent connection destination is a RAP-processing listener or a DCCM3 logical terminal defined in the DCCLTRAPHOST client environment definition.
- dc_rpc_call_s function
- dc_trn_begin_s function
However, if you execute function number 2 for a DCCM3 logical terminal, an error is returned.
When you specify Y in this definition, you need not execute the dc_clt_connect_s function. In addition, the dc_clt_disconnect_s function need not be executed since the permanent connection is automatically released when the dc_rpc_close_s function is executed.
- DCCLTTRSTATISITEM=statistics-item[,statistics-item]...
Specify the string constant that specifies the transaction branch statistics to be acquired. This definition is valid only when a transaction is started from the CUP.
- nothing
- Does not acquire statistics.
- base
- Acquires the following information as basic information.
- Transaction branch's identifier
- Result of the transaction branch's settlement
- Type of the transaction branch's execution process
- Name of the transaction branch's execution server
- Name of the transaction branch's execution service
- executiontime
- Acquires basic information and execution time information about a transaction branch.
- cputime
- Acquires basic information and CPU time information about a transaction branch.
You can specify nothing only once. Another statistics item has priority over nothing.
When acquiring statistics about a transaction, specify one of the following.
- trn_tran_statistics=Y in the transaction service definition
- -s option in the trnstics command
If this operand is not specified, the system follows the specification of the trn_statistics_item operand in the client service definition. If the connection destination is a RAP-processing server, the process follows the specification of the trn_statistics_item operand in the RAP-processing listener service definition.
- DCCLTTROPTIITEM=transaction-optimization-item[,transaction-optimization-item]...
Specify, by the following character string, an optimization item for enhancing the performance of a global transaction consisting of two or more user servers. This definition is valid only when a transaction is started from the CUP.
- base
- Optimizes the entire synchronous point acquisition processing (preparation, commit and rollback processing). OpenTP1 transaction control is performed in a two-phase commit way. Commit control between two transaction branches, therefore, requires four cycles of inter-process communication.
- If all of the following requirements are satisfied, the parent transaction branch performs commit processing instead of child transaction branches to reduce four cycles of inter-process communication required for commit control.
- - The parent transaction branch and child transaction branches are located under same OpenTP1.
- - The parent transaction branch uses a synchronous response RPC to call child transaction branches.
- - The object for the XA interface for a resource manager accessed by a child transaction branch is also linked to the parent transaction branch.
- asyncprepare
- Optimizes prepare processing if the system cannot optimize all the synchronous point acquisition processing because requirements on base are not satisfied. If all of the following requirements are satisfied, the system performs preparation before an RPC return when a child transaction branch executes a service request using an RPC issued from the parent transaction branch. The result is that two cycles of inter-process communication are reduced.
- Processing cannot be optimized using base. The parent transaction branch uses a synchronous response type PRC to call a child transaction.
- This optimization, however, delays the response time of a synchronous response type of RPC issued by the parent transaction branch. For a child transaction branch, this optimization increases the interval from prepare processing to commit processing. (In this status, a transaction cannot be settled if directions are not given from the parent transaction branch.) If OpenTP1 for the parent transaction branch is out of order, the communication between transaction branches impossible. Therefore, that lack of communication between those branches delays the validation of a journal file swap and check point dump file. The result may be that OpenTP1 for a child transaction branch is also out of order.
A duplicate transaction optimization item can be specified. Base, however, has priority over asyncprepare.
If this operand is not specified, the system follows the specification of the trn_optimum_item operand in the client service definition. If the connection destination is a RAP-processing server, the process follows the specification of the trn_optimum_item operand in the RAP-processing listener service definition.
- DCCLTTRWATCHTIME=maximum-communication-wait-time-during-transaction-synchronous-point-processing ~ <unsigned integer> ((1-65535)) (units: seconds)
Specify the maximum reception wait time for the communication between transaction branches (prepare, commit, rollback directions or response) during transaction synchronous point processing. This definition is valid only when a transaction is started from the CUP.
If the system does not give directions or response within a specified period, the transaction branch is rolled back if the first phase of two-phase commit is being processed. After the completion of the first phase, the system retries transaction settlement processing in a system process for a transaction service.
If this operand is not specified, the system follows the specification of the trn_watch_time operand in the client service definition. If the connection destination is a RAP-processing server, the process follows the specification of the trn_watch_time operand in the RAP-processing listener service definition.
- DCCLTTRRBINFO=no|self|remote|all
Specify whether to save information about the cause of a rollback as logs if a transaction branch has been rolled back. This definition is valid only when a transaction is started from the CUP.
- no
- Does not obtain rollback information.
- self
- Obtains rollback information in a log. The rollback information is only for the transaction branch that caused the rollback.
- remote
- Obtains rollback information in a log. The rollback information is the same information as self, plus the rollback information for any transaction branch that requested a rollback from another node's transaction branch.
- all
- Obtains rollback information in a log. The rollback information is the same information as remote, plus the rollback information for any transaction branch that requested a rollback from the local node's transaction branch.
If this operand is not specified, the system follows the specification of the trn_rollback_information_put operand in the client service definition. If the connection destination is a RAP-processing server, the process follows the specification of the trn_rollback_information_put operand in the RAP-processing listener service definition.
- DCCLTTRLIMITTIME=maximum-transaction-branch-execution-time ~ <unsigned integer> ((0-65535)) (units: seconds)
Specify the maximum transaction branch execution time. This definition is valid only when a transaction is started from the CUP.
The system automatically sets the expiration time of communication in synchronous point processing and of the dc_rpc_call_s function. This automatic setting prevents the period from the start of a transaction branch to the termination of synchronous point processing from exceeding a period specified in this operand.
- Time-out period of the dc_rpc_call_s function
If K is equal to or greater than a period specified in this operand, the system does not perform request processing, and returns as if a timeout error occurred.
If K is smaller than a period specified in this operand and if W is equal or less than a period specified in this operand minus K, the system adopts W as the time-out period.
If K is less than a period specified in this operand and if W is greater than a period specified in this operand minus K, the system adopts the period specified in this operand minus K as the time-out period.
K: Current time minus transaction branch start time
W: Time specified in the DCWATCHTIM operand
- Time-out period of communication in synchronous point processing
If K is equal to or greater than a period specified in this operand, the expiration time is one second.
If K is less than a period specified in this operand and if W is equal or less than a period specified in this operand minus K, the system adopts W as the time-out period.
If K is less than a period specified in this operand and if W is greater than a specified period in this operand minus K, the system adopts the period specified in this operand minus K as the time-out period.
K: Current time minus transaction branch start time
W: Time specified in the DCCLTTRWATCHTIME operand (or, if the DCCLTTRWATCHTIME operand is omitted, the time specified in the DCWATCHTIM operand)
If processing other than the above reception wait takes much time, a transaction branch may not be terminated within a period specified in this operand.
If the period specified in this operand has passed before the start of synchronous point processing, the transaction is rolled back.
If you specify 0, the system does not monitor the time.
If this operand is not specified, the system follows the specification of the trn_limit_time operand in the client service definition. If the connection destination is a RAP-processing server, the process follows the specification of the trn_limit_time operand in the RAP-processing listener service definition.
- DCCLTTRRBRCV=Y|N
Specify whether to receive notice of the completion of a rollback after sending rollback directions to a destination transaction branch of an RPC. This definition is valid only when you start a transaction from the CUP.
Y: Receives rollback completion notification.
N: Does not receive rollback completion notification.
If you specify N, the system terminates the local transaction branch without receiving notice of the completion of a rollback from a destination transaction branch of an RPC (without waiting for the completion of rollback processing in a destination transaction branch of an RPC).
If this operand is not specified, the system follows the specification of the trn_rollback_response_receive operand in the client service definition. If the connection destination is a RAP-processing server, the process follows the specification of the trn_rollback_response_receive operand in the RAP-processing listener service definition.
- DCCLTTRRECOVERYTYPE=type1|type2|type3
Specify a method for processing a transaction synchronous point when a UAP incurs an error. This definition is valid only when a transaction is started from the CUP.
If the time-out of an RPC occurs and the address of a process for an RPC destination is not determined or if a UAP for executing a transaction gets out of order, a transaction branch may not smoothly communicate with another transaction. In this case, therefore, transaction settlement may take a long time.
If the following failures occur, the system selects one of the following three methods of transaction synchronous point processing.
- type1
- Failure 1: Time-out of an RPC
- In this case, the RPC source transaction branch cannot determine the process in which a service request is being executed. It cannot, therefore, send a transaction synchronous point message to the RPC source transaction branch. Both of the RPC source transaction branch and the RPC destination transaction branch wait for a transaction synchronous point message. The result is that transaction settlement takes a long time.
- type2
- Failure 2: Malfunction of the RPC source UAP before reception of an RPC response
- In this case, the RPC source transaction branch cannot determine the process executing a service. It cannot, therefore, send a transaction synchronous point message to the RPC source transaction branch. The RPC destination transaction branch waits for a transaction synchronous point message. The result is that transaction settlement takes a long time.
- type3
- Failure 3: Nearly concurrent malfunction of the RPC source and destination UAPs after reception of a response from the RPC source UAP
- In this case, the transaction recovery process that has inherited each transaction branch is not notified that the remote UAP process is out of order. It, therefore, sends a transaction synchronous point message to a UAP process that does not exist. The result is that transaction settlement may take a long time.
In the following cases, transaction settlement may take a long time even if this operand is assigned type2 or type3.
- During the execution of an RPC, the status of the RPC destination UAP was changed (because of an increase in the load, the termination of the UAP, shutdown or the like). A service request has been re-transferred to the same UAP in another node.
- The version of destination OpenTP1 does not support this option.
- The destination transaction branch is occupied by processing other than processing for transaction synchronous point message reception.
If this operand is not specified, the system follows the specification of the trn_partial_recovery_type operand in the client service definition. If the connection destination is a RAP-processing server, the process follows the specification of the trn_partial_recovery_type operand in the RAP-processing listener service definition.
- DCWATCHTIMRPCINHERIT=Y|N ~ <<N>>
Specify whether the server should inherit the maximum wait time for a CUP response. By inheriting the maximum wait time for a CUP response, the server can be prevented from executing a service if the time-out of the CUP occurs.
Y: The server inherits the maximum wait time for a CUP response.
N: The server does not inherit the maximum wait time for a CUP response.
- DCSYSWATCHTIM=maximum-OpenTP1-control-response-wait-time ~ <unsigned integer> ((0-65535)) <<maximum-response-wait-time>> (units: seconds)
Specify the maximum value for the wait time from the sending of a request to the return of a response when under the control of OpenTP1. If a response is not returned within a specified period, the system returns an error message to the CUP.
If you specify 0, the system waits endlessly for the return of a response. In the default mode, the system applies DCWATCHTIM to the client environment definition.
You cannot dynamically change the maximum wait time for the OpenTP1 control response.
- DCCLTAUTHENT=Y|N ~ <<Y>>
Specify whether to authenticate the user for executing the dc_clt_cltin_s function.
Y: Authenticates the user.
N: Suppresses user authentication.
- DCCLTCONNECTINF=terminal-identification-information
Specify the terminal identification information. If you want to use hexadecimal numbers to specify this information, add 0x at the beginning of the information and use up to 128 digits (excluding the beginning 0x). If you use a character string, you can specify up to 64 characters.
When you use a permanent connection to communicate with a DCCM3 logical terminal, use EBCDIK code to specify the logical terminal name of the DCCM3 logical terminal as the terminal identification information. However, DCCM3 only validates the first 8 bytes (the 9th and later bytes are ignored).
The terminal identification information specified in this definition is referenced by the dc_rpc_open_s function. These functions then report that information to the DCCM3 logical terminal.
If you omit this definition, terminal identification information is not reported to the DCCM3 logical terminal.
However, if you execute the dc_clt_set_connect_inf_s function, the terminal identification information specified in this function is reported to the DCCM3 logical terminal when the dc_clt_connect_s function is executed.
This definition is valid when the host name and the port number of the DCCM3 logical terminal are specified in the DCCLTRAPHOST client environment definition and the dc_clt_connect_s function (specify DCNOFLAGS in the flags argument) is executed.
- DCSCDMULTI=Y|N ~<<N>>
Specify whether to use the multi-scheduler facility.
Y: Uses the multi-scheduler facility.
N: Does not use the multi-scheduler facility.
When you use the multi-scheduler facility, the system randomly selects one of multiple multi-scheduler daemons that are activated, to reduce the scheduling load.
When you specify Y for this operand, also see the description of DCSCDDIRECT, DCSCDPORT, and DCSCDMULTICOUNT in the client environment definition. This operand is invalid when the dc_rpc_call_to_s function is executed.
- DCSCDMULTICOUNT=number-of-multi-scheduler-daemon-processes ~<unsigned integer> ((1-4096)) <<1>>
Specify the number of multi-scheduler daemon processes. Specify the number of processes specified in the -m option in the scdmulti schedule service definition or less.
This operand is valid when DCSCDMULTI=Y, DCSCDDIRECT=Y, and DCSCDPORT are specified in the client environment definition. In this case, the system randomly selects one of the port numbers in the following range:
A to (A + B - 1)
where,
A: The port number specified in DCSCDPORT
B: The number of processes specified in DCSCDMULTICOUNT
- Lower limit: Port number value specified in DCSCDPORT of the client environment definition
- Upper limit: Lower limit value + the number of processes specified in DCSCDMULTICOUNT of the client environment definition - 1
- DCHOSTSELECT=Y|N ~<<N>>
Specify whether you want to have a gateway TP1/Server selected at random. This definition is valid only when multiple TP1/Servers are assigned as gateways.
Y: Selects TP1/Server to be used as the gateway at random.
N: Does not select TP1/Server to be used as the gateway at random.
When you specify Y, a gateway TP1/Server is selected at random from TP1/Servers that are specified at user authentication.
If an error occurs while the system is inquiring information from the name service of the gateway TP1/Server, TP1/Server where the error occurred is eliminated from the options. Then, the system selects another gateway TP1/Server at random in an attempt to switch.
When you specify N, gateway TP1/Servers are selected sequentially from the beginning of TP1/Servers that are specified at user authentication or those specified in the DCHOST client environment definition. If an error occurs while the system is inquiring information from the name service of the gateway TP1/Server, the system attempts to switch to the next specified TP1/Server.
When you specify Y in the DCSCDDIRECT client environment definition, the system attempts to switch the TP1/Server gateway if sending to the applicable port number fails.
- DCSCDLOADPRIORITY=Y|N ~<<N>>
Specify whether to distribute the load on the gateway TP1/Server that receives service requests with priority.
Y: Distributes the load on the gateway TP1/Server that receives service requests with priority.
N: Accepts the specification of the scd_this_node_first operand in the schedule service definition.
This definition is valid only when an RPC is executed without using the name service (specify Y in the DCSCDDIRECT client environment definition).
- DCCLTONLYTHISNODE=Y|N ~<N>
Specify whether to execute services on the specified node when the dc_rpc_call_to_s function is issued.
Y: Executes services only on the specified node.
N: First, attempts to execute services on the specified node. Depending on the service execution status, service requests may be transferred to another node.
- DCCLTNOSERVER=Y|N ~<<N>>
Specifies whether the environment being used is an environment in which TP1/Client never communicates with TP1/Server.
Y: In the environment being used, TP1/Client communicates with only a DCCM3 logical terminal and never communicates with TP1/Server.
N: In the environment being used, TP1/Client communicates with TP1/Server.
Always issue the dc_clt_cltin_s function even when Y is specified. In the logname argument, specify a value other than NULL. If you specify NULL, the dc_clt_cltin_s function returns a DCCLTER_INVALID_ARGS error. In the passwd argument, specify any value you like. You can specify NULL in the passwd argument.
- DCHOSTCHANGE=Y|N ~<<Y>>
Specify whether to switch the gateway TP1/Server to another TP1/Server if TP1/Client receives an error from the gateway TP1/Server in one of the following cases: (1) when a service request is issued (when the dc_rpc_call_s function is executed), (2) when the schedule service is being started, and (3) when the schedule service is being terminated.
Y: Switches the gateway TP1/Server.
N: Does not switch the gateway TP1/Server.
If this operand is incorrectly specified, TP1/Client assumes that Y is specified.
When N is specified, a service request (the dc_rpc_call_s function) immediately returns an error. If the schedule service is being started, a DCRPCER_OLTF_INITIALIZING error returns. If the schedule service is being ended, a DCRPCER_OLTF_NOT_UP error returns.
This definition is valid when more than one TP1/Server is specified with the target_host argument of the user authentication function and is specified with DCHOST in the client environment definition. This definition is invalid when a service request (the dc_rpc_call_s function) is executed during establishment of a permanent connection or within a transaction.
- DCCLTOPTION=client's-extension-option ~<unsigned hexadecimal integer> ((00000000-00000008)) <<00000000>>
Specify the option for extending the functionality of the client. To specify multiple options, specify the logical sum of the values of them. Even if you specify a value that cannot be specified, TP1/Client may operate incorrectly without generating an error code.
- 00000000
- When this option is specified, the functionality is not extended.
- 00000002
- When this option is specified, the client's timeout for awaiting a response (the value of DCSYSWATCHTIM or DCWATCHTIM in the client environment definition) is applied to reception processing of communication functions. This timeout is not applied every time a reception occurs. This timeout is applied only to the wait for a response from communication functions.
- Note that processing may be delayed if, for example, a fraction occurs during the decrement of time.
- 00000008
- When this option is specified, the host name length that TP1/Client can handle is extended from 63 characters to 255 characters.
If this operand is specified incorrectly, the KFCA02401-E message is output to the error log, and one of the following functions returns a DCCLTER_FATAL error.
- dc_clt_cltin_s function
- dc_clt_accept_notification_s function
- dc_clt_cancel_notification_s function
- dc_clt_open_notification_s function
- DCCLTNAMEXTEND=0|1 ~<<0>>
Increase the number of service information items acquired and managed by clients when multi-node servers are used.
0: Acquires a maximum of 128 service information items.
1: Acquires a maximum of 512 service information items.
Specify 1 for this operand when nam_service_extend=1 is specified in the name service definition of the TP1/Server that is used as gateway in a multi-node server configuration with 129 or more servers.
This operand is valid only when DCCLTLOADBALANCE=Y is specified in the client environment definition. When DCSCDDIRECT=Y is specified in the client environment definition, the specification of this operand is invalid.
- DCTRCPATH=trace-output-directory ~<character string>
Specify the absolute pathname of the directory for storing the error log files and trace files.
No files are output if the specified directory does not exist or the specification is invalid.
If this operand is omitted, the current directory is assumed.
- DCTRCERR=error-log-size ~<unsigned integer> ((0-1073741824)) <<4096>> (unit: bytes)
Specify the size of the error log files (dcerr1.trc and dcerr2.trc). These files are created in the directory specified with DCTRCPATH in the client environment definition, or in the directory where the CUP was executed.
If you specify 0 or an invalid value, or if there is no information to be output, then error log information is not output.
Although you can specify a maximum of 1 GB, when you have saved the error log trace in a file in an edited format, your machine may not be able to open such files, depending on the specifications of the machine. You must specify a valid value in accordance with the environment you are using.
- DCTRCUAP=UAP-trace-file-size ~<unsigned integer> ((4096-1073741824)) (unit: bytes)
Specify the size of the UAP trace files (dcuap1.trc and dcuap2.trc). These files are created in the directory specified with DCTRCPATH in the client environment definition, or in the directory where the CUP was executed.
If the specification is incorrect or omitted, or if there is no information to be output, UAP trace information is not output.
Although you can specify a maximum of 1 GB, when you have saved the UAP trace in a file in an edited format, your machine may not be able to open such files, depending on the specifications of the machine. You must specify a valid value in accordance with the environment you are using.
- DCTRCSOC=socket-trace-file-size ~<unsigned integer> ((4096-1073741824)) <<do not output>> (unit: bytes)
Specify the size of the socket trace files (dcsoc1.trc and dcsoc2.trc). These files are created in the directory specified with DCTRCPATH in the client environment definition, or in the directory where the CUP was executed.
If the specification is incorrect or omitted, or if there is no information to be output, socket trace information is not output.
Although you can specify a maximum of 1 GB, when you have saved the socket trace in a file in an edited format, your machine may not be able to open such files, depending on the specifications of the machine. You must specify a valid value in accordance with the environment you are using.
- DCTRCSOCSIZE=socket-trace-data-size ~<unsigned integer> ((64-4096)) <<256>> (unit: bytes)
Specify the data size of a socket trace to be output.
If the specification is incorrect or omitted, the default is validated.
- DCTRCMDL=module-trace-file-size ~<unsigned integer> ((4096-1073741824)) <<do not output>> (unit: bytes)
Specify the size of the module trace files (dcmdl1.trc and dcmdl2.trc). These files are created in the directory specified with DCTRCPATH in the client environment definition, or in the directory where the CUP was executed.
If the specification is incorrect or omitted, or if there is no information to be output, module trace information is not output.
Although you can specify a maximum of 1 GB, when you have saved the module trace in a file in an edited format, your machine may not be able to open such files, depending on the specifications of the machine. You must specify a valid value in accordance with the environment you are using.
- DCCLTPRFINFOSEND=Y|N ~<<N>>
Specify whether to send the identification information for a performance verification trace to TP1/Server.
Y: The identification information for a performance verification trace is sent to TP1/Server.
N: The identification information for a performance verification trace is not sent to TP1/Server.
When Y is specified for this operand and DCTRCUAP in the client environment definition is set to include performance verification trace information in the UAP trace, performance verification trace information is included in the UAP trace. If the identification information for the performance verification trace is sent to TP1/Server and the UAP trace is acquired, you can check the TP1/Client function execution time against the TP1/Server service execution time. In addition, you can also determine the progress of processing.
- DCCLTRPCMAXMSGSIZE=maximum-RPC-message-length~<unsigned integer> ((1-8)) <<1>> (unit: megabytes)
Specify the maximum length of user data that can be sent and received by using an RPC or the facility for receiving one-way messages from the server. This definition is valid when the following functions and arguments are specified.
Function |
Argument |
Input parameter length |
Output parameter length |
dc_rpc_call_s function |
in_len |
out_len |
dc_rpc_call_to_s function |
in_len |
out_len |
dc_clt_accept_notification_s function |
-- |
inf_len |
dc_clt_cancel_notification_s function |
inf_len |
-- |
dc_clt_chained_accept_notification_s function |
-- |
inf_len |
- Legend:
- -: Not applicable
If you specify 2 or a larger value in this operand, the maximum length of user data that can be sent and received is the value of this operand x 1024 x 1024 (bytes), rather than the value of DCRPC_MAX_MESSAGE_SIZE. If the specified value of an argument of one of the above functions is larger than the value of this operand x 1024 x 1024, the function returns a DCRPCER_INVALID_ARGS or DCRPCER_MESSAGE_TOO_BIG error.
If the length of received user data is larger than the value of the inf_len argument, the dc_clt_accept_notification_s function or dc_clt_chained_accept_notification_s function of the facility for receiving one-way messages returns a DCCLTER_INF_TOO_BIG error.
When specifying 2 or a larger value in this operand to send or receive user data larger than the value of DCRPC_MAX_MESSAGE_SIZE, note the following:
- You must also specify an appropriate value for the rpc_max_message_size operand in the system common definition for the TP1/Server to which you want to send a service request. Specify the same value for all the nodes specified for the all_node operand in the system common definition.
- If a request for a service is made to TP1/Server that does not support the rpc_max_message_size operand in the system common definition, operation is not guaranteed. If you execute either of the following RPCs, TP1/Server may go down:
Specify Y for DCSCDDIRECT in the client environment definition and issue the dc_rpc_call_s function.
Specify N for DCCLTONLYTHISNODE in the client environment definition and issue the dc_rpc_call_to_s function.
- If the response data length specified in the service function is larger than the value of rpc_max_message_size in the system definition x 1024 x 1024, the dc_rpc_call_s function or dc_rpc_call_to_s function returns a DCRPCER_INVALID_REPLY error.
- If the TP1/Server inter-node load-balancing facility transfers a service request to another node, a DCRPCER_NET_DOWN error may be returned.
- When using the permanent connection establishment function or transaction control function, you must specify an appropriate value for the rpc_max_message_size operand in the system common definition. If you do not specify an appropriate value, the dc_rpc_call_s function returns a DCRPCER_MESSAGE_TOO_BIG or DCRPCER_INVALID_ARGS error.
- If the service is not running on TP1/Server that supports the rpc_max_message_size operand in the system common definition, the dc_rpc_call_s function or dc_rpc_call_to_s function returns a DCRPCER_NO_SUCH_SERVICE_GROUP or DCRPCER_TRNCHK error.
- DCCLTRECVBUFSIZE=TCP/IP-receive-buffer-size ~<unsigned integer> ((8192-1048576)) (unit: bytes)
Specify the size of the TCP/IP receive buffer secured for each connection. Adjusting this value in relation to the buffer length of the remote system can improve the effectiveness of communication.
- Note:
- Upon receiving data, TCP returns a delivery confirmation packet (ACK).
- However, if the amount of received data is less than the size of the receive buffer, TCP does not immediately return the ACK (the ACK is delayed).
- In a communication environment in which the value of this operand is large and a small amount of data is sent and received, a delayed ACK might degrade performance. For details about delayed ACKs, see the TCP/IP documentation.
- Make sure that the value of this operand is less than the size of the TCP/IP receive buffer available in the OS.
If this operand is omitted, the system default is used.
- DCCLTSENDBUFSIZE=TCP/IP-send-buffer-size ~<unsigned integer> ((8192-1048576)) (unit: bytes)
Specify the size of the TCP/IP send buffer secured for each connection. Adjusting this value in relation to the buffer length of the remote system can improve the effectiveness of communication.
- Note:
- Make sure that the value of this operand is less than the size of the TCP/IP send buffer available in the OS.
If this operand is omitted, the system default is used.
- DCCLTTCPNODELAY=Y|N ~<<N>>
Specify whether to disable the Nagle algorithm.
Y: The Nagle algorithm is disabled.
N: The Nagle algorithm is not disabled.
Specifying Y for this operand might degrade sending efficiency during INET domain communication, increasing network load. Before specifying Y for this operand, carefully consider whether it is necessary to disable the Nagle algorithm in light of DCCLTSENDBUFSIZE and DCCLTRECVBUFSIZE operands in the client environment definition, and the network bandwidth.
For details about the Nagle algorithm, see the TCP/IP documentation.
All Rights Reserved. Copyright (C) 2006, 2009, Hitachi, Ltd.