Table of Contents
This chapter describes how to configure the TCP/IP gateway environment.
TCPGW compiles TCPGW library and implementations of custom.c and custom.h. To configure the TCPGW server, the following files must exist in each directory according to the operating system.
UNIX
| Directory | File Name |
|---|---|
| lib | libtcpgw.a, libtcpgw.so, libtmaxgw.a, libtmaxgw.so |
| lib64 | libtcpgw.a, libtcpgw.so, libtmaxgw.a, libtmaxgw.so |
Windows
| Directory | File Name |
|---|---|
| bin | tcpgw.dll, tmaxgw.dll |
| lib | tcpgw.lib, tmaxgw.lib |
TCPGW libraries are created under each directory when Tmax is installed. If custom.c or custom.h are needed, contact technical support.
To use TCPGW, TCPGW must be configured as a server in the Tmax environment file. Configuration is similar to the USC server among Tmax servers except that SVRTYPE is CUSTOM_GATEWAY instead of UCS. For TCPGW, only the SERVER and SERVICE sections have to be configured.
The following is an example of a Tmax environment file.
*DOMAIN
tmax SHMKEY=88000,
MINCLH=1,
MAXCLH=1,
TPORTNO=8800
*NODE
tmax1 TMAXDIR="/home/tmax",
APPDIR="/home/tmax/appbin"
*SVRGROUP
svg1 NODENAME=tmax1
*SERVER
tcpgw SVGNAME=svg1,
MIN=1,
MAX=1,
CPC=10,
SVRTYPE=CUSTOM_GATEWAY,
CLOPT="-- -P 5050 -N 10 -k 71673"
*SERVICE
TCPGW1
TCPGW2 SVRNAME=tcpgw,
SVCTIME=20,
SVRNAME=tcpgw
MIN
Number of processes of TCPGW. If the item value is larger than 1, then a different port number is used depending on the mode type.
| Type | Description |
|---|---|
| Server Mode | If set to 2 or more, each process must use a different listener port. Therefore, TCPGW uses its server count with the port number specified by the [–P] option of the CLOPT item (TCPGW option). For example, if the port number is 5050, then the first process uses 5050 and the second process uses 5051. The number is incremented by 1 for each additional process. When there are 2 or more processes, check if there are enough available port numbers for the number of specified processes. |
| Client Mode | A single port number is in standby on the remote node, and all TCPGW processes connects to the port. Following 3 methods can be used in the client mode:
|
CPC
Number of channels between the Tmax engine and TCPGW. When a Tmax client initiates a service call to the remote node, the CPC count must be set to the number of concurrent requests. However, for a non-blocking service, an appropriate value can be specified instead of the number of concurrent requests. Not many channels are required when a remote node initiates the request since TCPGW uses the tpacall of the Tmax engine.
CLOPT
Refer to the CLOPT Item (TCPGW Option) description.
SVCTIME
TCPGW can specify multiple services in a single server process and apply a different SVCTIME value for each service. If multiple services are specified in the SERVICE section each with a different SVCTIME value, then TCPGW uses the specified SVCTIME values.
For detailed information on configuration items of the SERVER and SERVICE sections, refer to "Chapter 3. Environment Configuration" in "Tmax Administrator's Guide".
Since the items that can be registered in the Tmax configuration file for TCPGW are limited, some options must be configured in the CLOPT item. The following options that define the operations of TCPGW must be well understood.
The following describes the TCPGW options.
IP address to listen when TCPGW runs in the server mode. This option is only used when connection is accepted only for the specified IP address.
When TCPGW runs in the client mode, the [-M] option can be used to modify the operation mode and specify the remote node IP to connect to.
In the following example, TCPGW connects to 10 incoming channels through 8080 port on the IP address 1.1.1.1 and 10 outgoing channels through 8081 port on the IP address 2.2.2.2. When running as a client, a separate IP address can be specified for the remote node as in the following.
-R 1.1.1.1 –P 8080 –r 2.2.2.2 –p 8081 –M 1 –N 10 –n 10
To listen using the IPv6 protocol, configure -XSERVER_IPV6 to 'Y'.
-R 2011::100:100 –P 8080 -XSERVER_IPV6=Y –r 2.2.2.2 –p 8081 –M 1 –N 10 –n 10
IP address of the remote node when TCPGW runs in the client mode.
Syntax has been enhanced as follows in order to configure multi-server channel information and backup functions in the client mode.
-r mainaddr[:mainport][/backupaddr[:backupport]][(type)][,...]mainaddr
Backupaddr is the IP address and mainport and backupport are the port numbers.
When using the IPv6 protocol, add the option, '-X CLIENT_IPV6=Y'. When configuring the mainaddr and backupaddr in the IPv6 protocol environment, surround the address with square brackets ([ ]) to distinguish the host address and the port number.
-r [2011::100:100]:8080(A)/[2011::100:200]:9090(A) -X CLIENT_IPV6=Y
Type can be configured as follows.
| Configuration Item | Description |
|---|---|
| I(IN CHANNEL) | Creates as many IN CHANNELS as specified by the [-N] option. |
| O(OUT CHANNEL) | Creates as many OUT CHANNELS as specified by the [-n] option. |
| A(ANY CHANNEL) | Creates as many IN CHANNELS and OUT CHANNELS as specified by the [-N] option and [-n] option, respectively. |
Port number to listen to in the server mode. Use a port number that is not used by other processes.
When in the client mode, it can also be used as a client port number as described for the [-R] option.
Port number where the remote node is in standby. Use a port number that is not in standby for other processes.
Option to change the port number and IP address specified in the server mode to the client mode.
The values specified using the [-R] and [–P] options are the IP address and port number for the server mode. This option changes them to the remote node.
"-M 0" is the basic mode of TCPGW. But if it is configured to "-M 1", then the mode changes from the server mode to the client mode. (Default Value: 0)
Option that converts data when the encoding system of the remote node is different from TCPGW.
Sent/received data are converted between ASCII, EBCDIC, Korean composition Hangul, and euc-kr characters. A data map file must be registered to use this option.
Refer to "1.3.5. Code Conversion" for code conversion. (Default value: not converted)
Number of incoming channels from the remote node to TCPGW.
This option is used regardless of the mode type.
Number of outgoing channels from TCPGW to the remote node.
This option only matters when TCPGW runs as a client.
Shared memory key value used to save the status of the channel specified with the aforementioned [-N] and [–n] options to enable the user to secure shared memory.
TCPGW doesn't secure shared memory beforehand by using this option but sends it to the user to use as needed. Generally, when TCPGW is started for the first time, this option is specified when calling the user routine (init_remote_info) and is used by the user to secure shared memory.
Special purpose CLOPT option added to the environment file without specifying the aforementioned IP address and port number.
Used for special purposes only. Cannot be used by general users.
Option for when TCPGW is used as a client. While [–r] and [–p] options are used to connect to a single remote IP, this option is used to connect to multiple remote nodes. This option is set to the file name that contains the connection information of the remote nodes which TCPGW uses to connect to the remote nodes.
Cannot be used with [-r], [-p] options or when TCPGW is used as a server. It can only be used in the client mode.
To use this option, connection must be maintained with the remote node. Therefore, the [-E] option, which connects to the remote node for every request, cannot be used.
When the remote node uses the IPv6 protocol, the ipaddr must be set in the IPv6 format and the ipv6 must be set to 'Y' to create and connect to the socket using the IPv6 protocol. The ipv6 configuration specified in this file takes precedence over the "-X CLIENT_IPV6" setting specified in the CLOPT section.
Port number that connects to the remote node when 2 or more TCPGW processes are executed while TCPGW is running in the client mode. If this option is specified, then TCPGW connects to the remote node using the port number and the server count values specified with the [-p] option.
For example, if the option is specified as "-p 8080 –C" and TCPGW is configured as MIN=3, then the first TCPGW process uses 8080 port number, the second TCPGW process uses 8081 port number, and the third process uses the 8082 port number.
Used when there are separate incoming and outgoing channels for TCPGW for communicating with the remote node, or when a request is sent from Tmax to the remote node.
When this option is specified, TCPGW sends the request from Tmax to the outgoing channel.
When data is received through the outgoing channel of the remote node, it is processed as the ACK reply of the previously sent data and the reply for the request is received through the incoming channel. This option is used when an ACK data, which indicates that the requested data has been transmitted successfully between TCPGW and the remote node, is sent and received. An ACK is not sent for a request message.
Service name used by TCPGW to request in the non-blocking or asynchronous mode.
When the non-blocking mode is used, Tmax's sending service tpforwards the request to TCPGW. Then TCPGW registers the service name to tprelay the reply received from the remote node to the receiving service. In the asynchronous mode, the Tmax client or service requests a service to TCPGW (tpacall with TPNOREPLY) and then shuts down. Later, the reply is usually received via TPNOREPLY from the remote node and is discarded. However, if the service name is specified, TCPGW sends the reply to the specified service via tpacall. The specified service name must be a registered service name.
Used when the user information is exchanged between a Tmax client or service and TCPGW. Data specified in this option is temporarily stored in TCPGW instead of being sent to the remote node. When the reply is received, the reply is attached to the user information and sent to the originating service.
Usually in the non-blocking mode, which uses separate services for sending and receiving, this option is used when data is sent from the sending service to the receiving service.
Same as the [-H] option. The [-H] option is used for both synchronous and asynchronous modes, but this option is only used when calling tpforward via tprelay. (Default Value: 0)
Method used to select the channel that Tmax service will use to send the reply to the remote node. When this option is specified, the reply must be sent to the remote node through the specified channel. If the option is not specified, then an available channel among those connected to the remote node will be selected. For future expansion, an arbitrary value must be specified.
For example, use an option of the form '-d 0' and the currently configured value is ignored.
Used when connecting and disconnecting channels for every request when TCPGW is running in the client mode.
If this option is used, the remote node cannot initiate a service request since there is no channel connected to the remote node. Furthermore, connecting and disconnecting the channel every time results in reduced processing speed. If the '-X OUTBOUND_REQ' option is used when connection failure occurs due to an outbound request, outmsg_recovery is called and the TPESYSTEM error is returned.
When the return data in get_msg_info() is abnormal, it returns -1. In this case if the [-e] option in the CLOPT section of the environment configuration file is configured, the channel shuts down. (TCPGW 3.9 version or later).
Option to use TCPGW in the asynchronous mode.
In the asynchronous mode, a service shuts down after making a request. A service reply is treated as a new request.
User-specified channel to use when a Tmax client or service requests a service via TCPGW to the remote node.
To use this option, the [-k] option must be used to store the status information of the channel. If the specified channel is invalid, then TCPGW returns an error.
There could be many communication equipment or a firewall between the remote node and TCPGW. In this case, if the channel in specific equipment or a firewall is not used for a specific period of time, then the channel can be disconnected unilaterally.
When the channel is disconnected, usually the TCPGW or remote node process detects this and reconnects to the channel, but there are times when the disconnection is not detected. This option is used to disconnect and reconnect the channel if the channel is not used during the specified time in TCPGW.
This option specifies the recovery service to allow cancellation of the previously processed service when the result cannot be sent to the remote node after the service is normally processed on Tmax.
If the remote node and session are both disconnected after the remote node calls a Tmax service, the service result is not sent to the remote node and is discarded. This can create a problem since the service is processed normally but the remote node recognizes it as an error.
If the service name is specified in this option, the reply is sent to the specified service instead when the gateway cannot send the reply to the remote node.
When TCPGW runs in the client mode, TCPGW first connects with the remote node. If the option is not configured when the connection disconnects from the remote node or there is no connection, then reconnection is attempted in 3 second intervals or when there is a service request. System load occurs from frequent attempts to connect to the remote node.
To resolve the system load problem, configure the time interval (s) to connect to the remote node and then specify the time with the [–t] option. TCPGW attempts to connect to the remote node using this time interval.
This option is required when a bit stream or a special character that indicates the end of data exists when receiving data from the remote node.
When the option is specified, then chk_end_msg is called when receiving data.
Enables the channel error detection (PING/PONG in the TCP level) function.
The following describes the options.
| Value | Description |
|---|---|
| 0 | When error recovery is detected on the main channel, then switch to the main channel occurs immediately. (default value) |
| 1 | Even if the main channel recovery is detected, switch to the main channel occurs at a safe point in time (when the channel is not in use). |
These values are used with the set_ping_msg, chk_pong_msg, and reset_ping_msg functions.
If the channel error monitoring function is enabled using the [x] option, then specify the maximum size of the message body that can be used when set_extping_msg is called. (Default Value : 0, maximum value : 268435455)
When the ping message is sent, the information of the sending message can be modified using the reset_ping_msg function. The message body size can be specified within the maximum size limit specified in this option.
This option specifies the maximum message number (sequence number) that can be used as UID. (Default Value: 50000, Maximum Allowed Value: 268435455)
The specified value is incremented by 1. Hence, the default value range is between 1 and 500001.
TCPGW can be configured to connect and disconnect each time it communicates with the remote node which creates too many logs related to closing the connection. This option is used to avoid recording such logs.
The following is an example of the logs.
(E)SVR6102 remote socket closed (E)SVR6103 remote gateway closed : ipaddr (E)SVR0030 read error
This option specifies to run the Channel IRT function (IRT COUSIN) when TCPGW is configured in a COUSIN group.
When TCPGW is not connected to the remote node, requests are sent to another TCPGW configured in the COUSIN group.
When TCPGW runs in the client mode, this option specifies the connection retry count for when connection to the remote node fails.
When this option is not specified, by default, connection is attempted only once.
If an attempt to reconnect to the main address fails, then connection is attempted to the backup address as many times as the retry count.
This option specifies the maximum timeout for TCPGW to connect to the remote node when it runs in the client mode. (Default Value: 3, Unit: seconds)
This option specifies whether to use the IPv6 protocol for creating the listener socket when TCPGW runs in the server mode.
The following describes the options.
| Configuration Value | Description |
|---|---|
| N | Use IPv4 protocol. Same as if not set. |
| Y | Use IPv6 protocol. |
This option specifies whether to use the IPv6 protocol for creating the socket to connect to the remote node when TCPGW runs in the client mode.
The following describes the options.
| Configuration Value | Description |
|---|---|
| N | Use IPv4 protocol. Same as if not set. |
| Y | Use IPv6 protocol. |
The following is an example of the server/client TCPGW environment configuration.
*DOMAIN
...
*NODE
...
*SVRGROUP
...
*SERVER
testtcpgw SVGNAME = svg1,
MIN = 1,
MAX = 1,
CPC = 10,
SVRTYPE = CUSTOM_GATEWAY,
# For the client TCPGW
CLOPT = "-- -r 168.126.185.131 -p 5050 -n 5 -k 71673 "
# For the server TCPGW
#CLOPT = "-- -P 5050 -N 5 -k 71673"
*SERVICE
TESTTCPGW SVRNAME = testtcpgw
The previous environment configuration starts up a single TCPGW instance named testtcpgw when Tmax starts up. The service name of testtcpgw is TESTTCPGW and since the timeout value is not configured it waits infinitely.
For the client TCPGW, the remote node's IP address is "168.126.185.131" and the remote node's listen port number is "5050". When testtcpgw starts up, it creates 5 connections with the remote node. For the server TCPGW, it is configured to run as a server and the listen port number is "5050" and the remote node and testtcpgw creates 5 connections.
In each case, the specified [-k] option value is passed as an argument of the factor(init_remote_info()) function in custom.c, and is used to store necessary data in the shared memory.
After connection is made with the remote node, TCPGW can use both synchronous/asynchronous methods regardless of the server/client setting. Each method will be described regardless of the setting.
*DOMAIN
...
*NODE
...
*SVRGROUP
...
*SERVER
testtcpgw SVGNAME = svg1,
MIN = 1,
MAX = 1,
CPC = 10,
SVRTYPE = CUSTOM_GATEWAY,
#For the client TCPGW
CLOPT = "-- -r 168.126.185.131 -p 5050 -n 5 -k 71673 -d 0"
#For the server TCPGW
#CLOPT = "-- -P 5050 -N 5 -k 71673 –d 0"
*SERVICE
TESTTCPGW SVRNAME = testtcpgw, SVCTIME = 30
The previous environment configuration starts up a single TCPGW instance named testtcpgw when Tmax starts up.
The service name of testtcpgw is TESTTCPGW and the timeout is 30 seconds. If there is no reply within 30 seconds, then a timeout error is returned. Timeout only checks requests from a Tmax client or service to the remote node. Services requested from the remote node must specify the timeout setting in the request.
TCPGW uses the UID when a request is made to the remote node from a Tmax client or server. UID, the unique identifier of the message, must be passed to TCPGW so that the reply can be sent to the originating client/server.
When a service is requested from the remote node, TCPGW sends the reply to the requested channel. If the requested channel is disconnected, the data is abandoned.
Non-blocking TCPGW is generally used when a service is requested from a Tmax service to the remote node. A general Tmax client cannot directly call services in this mode. Only the synchronous or asynchronous method can be used directly in the Tmax client.
*DOMAIN
...
*NODE
...
*SVRGROUP
...
*SERVER
testtcpgw SVGNAME = svg1,
MIN = 1,
MAX = 1,
CPC = 5,
SVRTYPE = CUSTOM_GATEWAY,
#for the TCPGW client
CLOPT = "-- -r 168.126.185.131 -p 5050 -n 5 -k 71673 -S RECVSVC
-H 20 -d 0"
#For the server TCPGW
#CLOPT = "-- -P 5050 -N 5 -k 71673 -S RECVSVC -H 20 –d 0"
sendsvr MIN = 1, MAX = 1
secvsvr MIN = 1, MAX = 1
*SERVICE
TESTTCPGW SVRNAME = testtcpgw, SVCTIME = 30
SENDSVC SVRNAME = sendsvr
RECVSVC SVRNAME = recvsvr
The previous environment configuration starts up a single TCPGW instance named testtcpgw when Tmax starts up. The service name of testtcpgw is TESTTCPGW and the timeout is 30 seconds. If there is no reply within 30 seconds, then a timeout error is returned. If the remote node does not return a reply within the set time (30s), then TCPGW sends a timeout message to RECVSVC (tprelay). The RECVSVC service must check the tpurcode value and when the value is larger than 0 indicates an error which must be handled. The tpurcode is set and returned for all errors including timeout. For detailed information about errors, refer to "Appendix A. TCPGW Error Codes".
If a Tmax client or server initiates a call to the SENDSVC service, SENDSVC preprocesses the request and sends the control (tpforward) to the TESTTCPGW service. TCPGW first sends the channel cancellation message to the Tmax engine and requests the service to the remote node. When a reply is received, the result is sent (tprelay) to the service specified in the [-S] option. After RECVSVC processes the result and returns, the reply is sent to the originating client or service. Using the non-blocking mode, more requests can be processed using less number of service threads (SENDSVC). The UID is used to identify the request as in the blocking mode.
A service requested from the remote node can be specified as synchronous or asynchronous using the [-d] option, or setting TPNORELAY to the flags in the user defined function (get_msg_info).
In the former case, if the [-d] option is not used, the reply can be sent to another channel. In the latter case, the reply is not sent to the remote node. Even if the [-d] option is used, if TPNOREPLY is set in the flags item, then the service reply is not sent to the remote node.
*DOMAIN
...
*NODE
...
*SVRGROUP
...
*SERVER
testtcpgw SVGNAME = svg1,
MIN = 1,
MAX = 1,
CPC = 5,
SVRTYPE = CUSTOM_GATEWAY,
#Asynchronous
CLOPT = " -- -P 5050 -N 5 -k 71673"
#Synchronous
#CLOPT = "-- -P 5050 -N 5 -k 71673 –d 0"
*SERVICE
TESTTCPGW SVRNAME = testtcpgw
When a Tmax service is requested from the remote node, TCPGW uses tpacall and thus does not receive a reply. The number of allowed concurrent tapcalls is limited (default value:8). When there are 8 or more concurrent requests, the MAXSACALL count must be set to an appropriate number in the DOMAIN section.
Communication channel between TCPGW and the remote node usually requires complex communication equipment or a firewall. If the channel is not used for a period of time, then the communication equipment or the firewall disconnects the channel and blocks the communication.
To prevent this, a signal can be sent and received at regular intervals but this requires having to consider both TCPGW as well as the opposing program. Instead, the reconnection method is used to disconnect and then reconnect the channel if it is not used for a period of time.
Using the following configuration, TCPGW disconnects a channel that has not been used within 600 second (10 minute) interval.
*DOMAIN
...
*NODE
...
*SVRGROUP
...
*SERVER
testtcpgw SVGNAME = svg1,
MIN = 1,
MAX = 1,
CPC = 5,
SVRTYPE = CUSTOM_GATEWAY,
#Asynchronous
CLOPT = " -- -P 5050 -N 5 -k 71673 -w 600"
#Synchronous
#CLOPT = "-- -P 5050 -N 5 -k 71673 -w 600"
*SERVICE
TESTTCPGW SVRNAME = testtcpgw
The following is an example of configuring TCPGW client that connects with IP addresses of multiple nodes.
*DOMAIN
...
*NODE
...
*SVRGROUP
...
*SERVER
testtcpgw SVGNAME = svg1,
MIN = 4, #Set to the number of addresses (ip_addr) configured in the configuration file below.
MAX = 4,
CPC = 10,
SVRTYPE = CUSTOM_GATEWAY,
CLOPT = "-- -f config -n 5 -k 71673"
*SERVICE
TESTTCPGW SVRNAME=testtcpgw
The following is an example of the configuration file.
# tcpgw_no ip_addr port_no in_channel out_channel 0 192.168.1.1 7717 5 5 1 192.168.1.2 7727 5 5 2 192.168.1.3 7737 5 5 3 192.168.1.4 7747 5 5
The following is a configuration file for the IPv6 protocol.
# tcpgw_no ip_addr port_no in_channel out_channel ipv6 1 192.168.1.1 7717 5 5 2 www.tmax.co.kr 7727 5 5 N 3 2011::100:100 7737 5 5 Y 4 2011::100:200 7747 5 5 Y
The following describes the items in the configuration file.
| Configuration Item | Description |
|---|---|
| tcpgw_no | TCPGW process number that starts from 0. |
| ip_addr | Remote host name or IP address. |
| port_no | Remote port number. |
| in_channel | Input channel number to which the request is sent. |
| out_channel | Output channel number for the reply. |
| ipv6 |
|
The ip_addr, port_no, and ipv6 items of the configuration are created for both main and backup. When connection with the main configuration fails, then it connects with the backup configuration. When the backup setting is not configured, the same values are set to the main and backup.
# tcpgw_no ip_addr port_no in_channel out_channel ipv6 0 2011::100:100/192.168.1.3 7717 5 5 Y/N 0 2011::100:200/2011::200:200 7727 5 5 Y 1 192.168.1.1/192.168.1.2 7737/7738 5 5 1 www.tmax.co.kr 7747/7748 5 5 Y/N
TCPGW can configure the user header and use it in the server/client mode, and synchronous/asynchronous methods. The user header can only be used when a service is requested from the Tmax client or service to the remote node, but not in the reverse direction.
The data specified with the user header is not sent to the remote node but is stored in the temporary TCPGW using its UID. When a reply is received, it finds the UID in the reply data and sends the user header data for the UID with the reply data. The length of the user header data can be up to 256 bytes.
The following is an example of the user header environment configuration.
*DOMAIN
...
*NODE
...
*SVRGROUP
...
*SERVER
testtcpgw SVGNAME = svg1,
MIN = 1,
MAX = 1,
CPC = 10,
SVRTYPE = CUSTOM_GATEWAY,
#For the client TCPGW
CLOPT = "-- -r 168.126.185.131 -p 5050 -n 5 -H 9 -h 10"
#For the server TCPGW
#CLOPT = "-- -P 5050 -N 5 -H 9 -h 10"
*SERVICE
TESTTCPGW SVRNAME = testtcpgw
The user header can be used in any mode, and can be specified in two ways. The user header length can be set differently when using the tpforward method (non-blocking) and when using other services (blocking, asynchronous).
The [-H] option is used for specifying the user header for all service request types, while the [-h] option can only be used for the tpforward method (non-blocking).
In the previous environment configuration, 9 bytes of the transmitted data are stored for general services and10 bytes are stored for tpforward, and only the remaining data are sent to the remote node. When the reply is received from the remote node, the stored user header is attached to the reply data and returned.
The user can selectively use the user header through various methods.
The following is a typical method of using the user header.
Specifying tprelay or the asynchronous service name
When it is used in the non-blocking method, the service to tprelay can be specified in the user header. When the service is also specified with the [-S] option, TCPGW first calls the user function (get_service_name) in the user header and finds the service. When the service is not found, the service specified with the [-S] option is used. When a call is made using asynchronous method via tpacall with TPNOREPLAY, the reply data is processed as described above.
Storing key data
In the non-blocking mode, there are separate sending and receiving service processes. The sending service processes tasks in the database and sends data to the remote node, and then terminates. This is often used when the receiving service needs to know the database key information that was processed by the sending service, or when the receiving service needs to store important information to rollback database changes when an error occurs on the remote node.
Channels can be connected and disconnected for every request by using the [-E] option. The backup channel can be specified in this situation. Use the [-r] option of the CLOPT item to specify the server and backup server's port number to configure the backup channel.
The following is an example of the configuration.
SERVER Name CLOPT = literalㅇ
The following is an example of configuring the CLOPT item.
CLOPT = "-- -r [ip]:[port]/[backup ip]:[backup port] –E"
When CRYPTPORT is specified in the NODE section, only client data accessed using the port set in CRYPTPORT is encrypted. For detailed description about the configuration items in the NODE section, refer to "Chapter 3. Environment Configuration" in "Tmax Administrator's Guide".
By using the set_ping_msg() and the chk_pong_msg() functions, the ping message is sent to periodically check the channel's error status. Since a ping message only has a header, it cannot be used to send data.
Instead of the set_ping_msg() and chk_pong_msg() functions, the set_extping_msg(), chk_extpong_msg(), and reset_extping_msg() functions can be used to transmit data. For detailed information about these functions, refer to "3.3. custom.c".
The following is an example of the configuration.
[DEFAULT :] SERVER Name [,CLOPT = -b len]
| Item | Description |
|---|---|
| CLOPT = -b len | Maximum size of ping message data. Optional. |
Configure the [-r] option in the CLOPT item as follows in order to add the ping message's data and send it.
CLOPT = “-- -r 192.168.1.31 –p 5060 –n 1 –x 1 –b 64”