Switching facility of TP1/Server as a gateway

If an error occurs in TP1/Server as a gateway, after the error is returned, the definition of TP1/Server as the gateway needs to be changed. However, if multiple TP1/Servers are specified as gateways, TP1/Client switches to the next specified TP1/Server if an error occurs.

When CUPs request service at the same time, inquires will become concentrated on the name service of a single gateway TP1/Server, thus increasing the load. The random selection of multiple TP1/Servers distributes the load on the gateway TP1/Server. For details on the load distribution by random selection of multiple TP1/Servers, see Subsection 2.3.13 Load distribution for TP1/Server as a gateway.

Organization of this subsection: (1) TP1/Server switches because of the following conditions.; (2) How to specify the switch destination OpenTP1; (3) Sequence of switching

(1) TP1/Server switches because of the following conditions.

If an error is detected in TP1/Server as a gateway when performing user authentication.
If the communication fails with the name service of TP1/Server, which is used as a gateway when issuing RPCs.
If the communication fails with the schedule service of TP1/Server, which is used as a gateway when issuing the RPC without the name service.
If the communication fails between the transaction accept service of TP1/Server when issuing the RPC that allocates the transaction management process.
If the client receives an error response from TP1/Server used as a gateway while a service is being requested, a schedule service is being started, or a schedule service is being terminated.

(2) How to specify the switch destination OpenTP1

Specify multiple TP1/Servers, which are used as gateways, in the target_host argument of the user authentication function.
Specify multiple TP1/Servers, which are used as gateways, in DCHOST of the client environment definition.
In addition to the above specifications, specify DCHOSTCHANGE=Y in the client environment definition to switch TP/Server when:
The client receives an error response from TP1/Server used as a gateway while a service is being requested, a schedule service is being started, or a schedule service is being terminated.

(3) Sequence of switching

When target_host is specified as an argument of the user authentication function, TP1/Servers switch in the sequence they are specified in target_host.
If target_host is not specified as an argument of the user authentication function, TP1/Servers switch in the sequence specified in DCHOST in the client environment definition.

2.3.12 Switching facility of TP1/Server as a gateway

(1) TP1/Server switches because of the following conditions.

(2) How to specify the switch destination OpenTP1

(3) Sequence of switching