When the Baan ERP went live for years, it’s time to worry about the data growing and cleaning. The administrators will sooner or later receive calls such as why the reports take ages to print out, why the backup takes longer and longer, etc. In Baan, there is an ability to shrink your live data and put to separate database or even separate servers – Archiving.

Good preparation always gives a good result. Here I am talking about how to create a separate / remote Oracle database for the archiving companies.

1) Create on the remote server an oracle instance (say ARCH) with tablespace-names: ARCHDAT and ARCHIDX.

2) Specify in the $BSE/lib/ora/storage_params the data and index tablespaces (ARCHDAT and ARCHIDX) for the archive company (say company 743):
*:743:T:PCTFREE 5 TABLESPACE ARCHDAT INITRANS 3 STORAGE(INITIAL 16k NEXT 1M PCTINCREASE 0)
*:743:I:PCTFREE 5 TABLESPACE ARCHIDX INITRANS 3 STORAGE(INITIAL 16k NEXT 1M PCTINCREASE 0)
*:*:T:PCTFREE 5 TABLESPACE DATA INITRANS 3 STORAGE(INITIAL 16k NEXT 1M PCTINCREASE 0)
*:*:I:PCTFREE 5 TABLESPACE INDEXES INITRANS 3 STORAGE(INITIAL 16k NEXT 1M PCTINCREASE 0)

3) Create a new database definition in baan session: ttaad4110m000
Say database 010 with parameters: ORACLE_HOME=d:ORACLEORA81,ORACLE_SERVICE_NAME=ARCH,NLS_SORT=BINARY
and add the service ARCH in the tnsnames.ora file

4) In session ttaad4111m000: assign all tables of company 743 to database 10 and do a convert to runtime.

5) Add databasegroup ‘baandb’ to database type 010

6) Add the baan user to database group 010
(remember the database-user and database-group and password must be the same for both databases, this is a limitation of Baan: the ora_user file can
contain only one entry for one baan user).
Now it should be possible to access company 743 on the remote database from the baan sessions.

Note:
i. When receiving error 402 (wrong setup) the national language character set is not right.
ii. If in both databases the charactersets are the same, then you can specify in file $BSE/lib/defaults/db_resource the parameter: nls_lang (example: nls_lang=american_america.we8iso8859p1). Otherwise NLS_LANG must be specified in $BSE/lib/tabledef6.x (x=1 stands for BaanIV, x=2 for BaanV and higher).
Charachterset information can be retrieved by the SQL-query:
select * from nls_database_parameters where parameter in (‘NLS_LANGUAGE’, ‘NLS_TERRITORY’, ‘NLS_CHARACTERSET’ );
iii. Use the latest portingset if possible.