9.39 DEFINE_OS_400_SERVER

...

The time taken to define a server is minimal.

Arguments

Return Values

Technical Notes

• It is very strongly recommended that server definition logic is coded in one and only one function, rather than scattered and repeated through many RDML functions. This approach will isolate your application from future changes to the server(s) that are being used.

• The LU partner name specified must be either:

• The fully qualified Partner LU (Logical Unit) name that has been previously defined to the Communications Router.
  
  This is usually formatted "Network Name.Control Point Name" and corresponds to the IBM i DSPNETA values of "Local Network ID" and "Local control point name".
  
  Normally any Communications Router that allows 5250 emulation sessions to work will have these details already configured ready for use.
  e.g. APPN.SYDASD25

• Or, if you have configured an "alias" name for the LU name it may be used in place of the fully qualified LU Name.
  e.g. 5250PLU

• The translation tables specified are actually uploaded from the IBM i server during a connect. The initial connect phase needs to translate the initial connection details via a default translation table as the tables you specify have not yet been uploaded.
  The initial connect details (sent to the server) include:

• The partition

• The language code

• The task identifier

• The client to server conversion table name

• The server to client conversion table name
  
  Use only English alphabetic characters A through Z and the numbers 0 through 9 in the names of the objects listed above to avoid potential problems in this area.

• It is recommended that you use SSN values that are meaningful to end users (e.g. CHICAGO, BOSTON, CHARLIE1, etc.) as they may appear in messages from time to time.

• SSN names must be unique and start with a character in the English language alphabet (uppercase A through Z).

• A server may be repeatedly defined and (re)defined when it is not connected. If you attempt to (re)define a server that is currently connected a fatal error will result.

• The details defined are not persistent. They are lost when the X_RUN command completes. You may choose to define your own set of SQL based tables to hold server details and actually read the table(s) to get values to be passed on to this Built-In Function.

• Please experiment with these facilities first and then design some server architecture for your organization that has these characteristics:

• It matches your organization's requirements.

• It is quick and easy to change.

• It is extendible.
  
  Do this before launching into any large scale design or development project.

• The client's date format will be automatically passed to the server. If the date formats are different (e.g. MDY vs DMY), the server will automatically return data in the client's format.
  
  The client's date format can be changed from the default by specifying the x_run parameter DATF=. Please refer to Standard X_RUN Parameters for more information about this parameter.
  
  Note that if the client and server date formats are different, Date format validation rules specifying exact formats (e.g. DDMMYY) will fail (as the data may be returned as MMDDYY). Date format SYSFMT is recommended where clients need to use different date formats (e.g. USA and UK clients).

Notes on Commitment Control

• If Start Commitment Control is Y, LANSA will automatically start and end commitment control. See User Exit F@BGNCMT - Start Commitment Control and User Exit F@ENDCMT - End Commitment Control in the LANSA for i User Guide for details.

• When the server has been indicated as having commitment control started, it will effect all subsequent COMMIT and ROLLBACK commands issued.
  When a COMMIT or ROLLBACK command is issued the routine involved loops through all currently connected servers.
  To each one that has commitment control active, it issues a "commit" or "rollback" request and then waits for the server to respond before proceeding.
  This is done after a commit/rollback has been issued correctly to the local/client database management system.

A Note on Error Handling

It is very strongly recommended that you avoid building complex error handling schemes into your applications. Use a very simple trap like this at all levels of your application:

     if (#retcode *ne OK)

...

         abort msgtxt('Failed to .............................')
     endif

Let the standard error handling Built-In Function to every generated application take care of the problem. Situations have arisen where user defined error handling logic has become so complex as to consume 40 - 50% of all RDML code (with no obvious benefit to the application). Do not fall into this trap.

DBCS Considerations

When the server is indicated as being DBCS capable an additional translation table must be locatable on the client PC.

...